British Astronomical Association
Supporting amateur astronomers since 1890

Secondary menu

Main menu

Home Forums Telescopes
Terms of use

USB over Ethernet

29 posts / 0 new
Last post
Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
USB over Ethernet

As I am in the process of planning a new observatory (probably a roll off shed) for our new home, I am looking at options for control of the scope. I had wondered about logging in remotely to a laptop left by the telescope to control it, but recently saw there are ethernet adaptors made by Startech that allow 4 USB sockets to be accessed directly via a Cat5/6 cable - perhaps as much as 100m long.

Now, that would save me a lot of palaver in that I could have the CCD, an autoguiding camera and mount all controlled down the same cable with perhaps enough capacity left over for an all sky camera, or more likely, an electronic focuser. 

Looking at the Amazon feedback I see mention that others have done this, so I wondered if anyone had a story to tell on this. Does it work seamlessly with imaging or do we end up with the USB connections interfering with each other and messing things up - I recall using USBs on a laptop where moving the USB mouse impacted the signal from a USB record turntable.

Thoughts anyone?

nickjames's picture
Offline
Last seen: 4 hours 11 min ago
Joined: 12/10/2013 - 18:16
Grant,

Grant,

My only experience of USB over Ethernet has been very bad, and that was using a pro system in work. Much easier to buy a cheap PC from ebay, stick it in the observatory, leave it running permanently, connect a Cat 6 LAN cable to it, then run all the USB connections locally. You can then either remote desktop to it or run ASCOM remote from inside. Much more reliable and it'll set you back £50 or so.

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Nick

Thanks for that. The only colleague I could find who had used one much in anger was merely feeding keyboard and mice down one Cat5. Unsurprisingly, he didnt have much problem. 

Offline
Last seen: 1 month 1 week ago
Joined: 05/03/2014 - 00:00
USB over Cat5/6 cabling and general hints.

Grant.

I used to use medical rated versions of the Startech USB over Cat5 cabling extenders in medical imaging systems and have some experience with them.  

In some circumstances they do work well but the achilles heel is that all the data from your four devices at the client end flows through a single USB 2 port at the host end and is therefore limited to 480Mbs split four ways. These devices work well where the total data throughput from all the client devices never reaches that 480Mbs limit and that all the devices are roughly balanced in data throughput.

Problems can occur where a modern large sensor CCD/CMOS camera hogs all the available bandwidth for thirty seconds or more while downloading to the host and regular polling of the mount by your observatory software can not occur, in some cases the mount will “time out“ and require a reboot to reestablish USB communication.

This problem with the main camera taking all the available bandwidth has been addressed to some extent with the latest manufacturers offerings that include an onboard memory buffer in the camera that allows for a reduced data rate on the USB link however at this time I have only seen this with USB 3 CMOS cameras.

It is also the case that you will most likely need many more than just 4 USB ports in the observatory as you will find as you progress along the automation route.

In my own observatory (dome) I have (up to, depending on OTA configuration) 12 USB ports in use, main camera, guide camera, fiter wheel, AO unit, mount, focuser, rotator, dew controller, dome controller, cloud/rain/wind monitor, UPS supply for the PC and observatory power switch controller.

Depending how far along the remote control/automation path you plan to go if you begin adding in plate solving and refined mount pointing/closed loop slewing plus an all sky camera, well, a single USB over Cat5 cabling extender just won‘t be sufficient, you could be looking at two or even three separate extenders and at that price you might as well use a dedicated laptop or PC in the observatory controlled remotely by TeamViewer or Microsoft Remote Desktop from a separate laptop/tablet/PC in the house.

In the past my observatory was controlled by a dedicated desktop PC, that blew up due to damp, the ATX power supplies of conventional desktop PC’s are not rated for use in a condensing atmosphere and one dreary damp November the computers ATX power supply let go and damaged several connected devices. This was replaced with a laptop powered by a weatherproofed 19v supply. The laptop worked well until the temperature fell below freezing where the RAM memory clocks would drift and the dreaded ‘Blue Screen of Death’ would appear as Windows gave up. That problem was resolved by placing the laptop in a ventilated cabinet on top of low voltage pet warmer pad that kept the laptop above freezing. Eventually when that laptop came to the end of life it was replaced by an industrial fanless sealed PC with specially rated components for use below freezing, this works perfectly but at a high price. This type of industrial PC was a priority for my observatory because it is run fully autonomously for much of the year with no one in attendance to step in and deal with failures.

From the house to the observatory I ran three Cat5 cables, two in use and one spare, a multicore cable that is used for the alarm and halon gas fire supression system and an armoured power cable connected back to the main house fuse box. One of the Cat5 cables connects to a network switch in the observatory, the other end to the house router. The second Cat5 cable connects to a second separate switch and a pair of surveillance CCTV cameras, one inside the oberservatory the other outside, both connected back to the main house alarm system. Wireless networking is just not reliable enough at my location, it was not an issue with distance so much as too many other users nearby using the same wireless channels plus a requirement by my house insurer that only wired CCTV networking was used as many ne'er do well’s carry wireless network jammers while engaged on their activities. Networking over the powerline was considered but a line interactive UPS in the observatory does not pass the signal downstream of the UPS and severely attenuates the signal upstream.

All the low voltage signal cables from house to observatory run in a single 90mm diameter plastic pipe, buried a metre below ground, with a couple of spare pull-through draw wires in case I need to add or replace in the future. The armoured mains cable runs in the same trench but buried a further 450 mm below the pipe carrying the signal cables to prevent mains bourne interference and comply with local planning rules where minimum depth for a buried mains cable was stipulated at 1.2mtrs below ground level. The mains supply is configured as a TT system, that is, only live and neutral leave the house protected by a dedicated RCD in the main house fuse box and a two metre earth rod was driven into the ground right next to the observatory connected to a single earth bus bar installed inside the observatory. All metal work, including the steel pier, and observatory device power supply earths run back to that single point to avoid possible earth loops. The observatory computer is protected and maintained by a line interactive UPS, the dome shutter is battery operated and its supply is recharged by a solar panel on the dome roof. In case of loss of communication between shutter and dome controller or PC failure the shutter closes automatically. As well as a cloud, wind and rain monitor talking to the observatory computer and initiating a shutter closure command, a separate, independent rain sensor on the observatory roof triggers a shutter closure irrespective of the observatory computer.

In use, the observatory is controlled by ACP software that I can log in to from anywhere via TeamViewer but apart from giving it a list of targets ACP is in control of all the observatory functions including turning off/on frost protection heaters and the dehumidifier, observatory startup and shutdown, weather monitoring and target sequencing, switching on the flats panel etc. At the same time I can log into the house alarm and security system via a browser and watch the live video stream from my security cameras. When a planned target observation is complete, however many hours, days, weeks or months that may take, ACP sends the folder of images over the internet to my dropbox account as well as an email notification for me to post process wherever I might be.

Some astronomers are reporting successful control of small observatories using ’micro’ or ’stick’ computers run ‘headless’, that is without monitor keyboard or mouse, mounted up on the OTA with an additional 12v powered USB hub and 12v power distribution bus. Usually these micro stick PC’s only have a couple of USB ports plus either a cabled network port or wireless LAN and Bluetooth. As long as the main camera is assigned one of the available USB ports alone while everthing else goes via the second USB port and hub they do seem to work. The slightly larger mini PC‘s usually have four USB 2 sockets and are still light enough to be mounted alongside the OTA for short cable runs. Communication and control is only via TeamViewer, Remote Desk Top, or similar, from another computer but with the ability to run a full version of Windows and very short USB cables to everything else bar the dome/roof which would probably need an active extension from the USB hub up on the OTA to reach. Though not very fast and small memory storage I have read they will run observatory software including plate solving quite successfully with a minimum of cabling running up the mount. If you are not planning a fully autonomous observatory and are happy to attend to switching heating/ventilation and dehumidifying on off plus flats panels as necessary and don’t require a cloud/rain/wind sensor or weather station then these little ’Micro Stick’ or Mini PC’s may offer you all you want.

Do remember to check with your insurance company, I found it impossible to get insurance cover for the observatory without providing the connection into the house alarm and CCTV systems, when you look carefully at the wording of many household policies high value items used outside the home are not covered if you are actually away from home and out of sight of the items while general outbuilding cover is limited to a fraction of the cost of a typical observatory, I also discovered that in the case you are underinsured the insurance company can refuse to payout anything in the case of a claim!

Rather a long reply but hopefully of some use.

William.

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
USB over Cat5/6

Thanks for the comprehensive reply. 

Fortunately, I am not going for a fully automated solution. In the absence of a shed that lends itself to automation I don't think its worth the hassle for me just to avoid going outside in the dark/cold for 10 minutes at the start and  the end of each session. Can see the joy and challenge in it though.

In light of what seems to be the opinion of USB over internet, I will probably go for a NUC hung on the back of the scope so all I need a is 5A 19V power supply cable and an ethernet cable connecting to them to Remote Assist/desktop via.

A thermostat/heater on the NUCs will be needed of course as few computer drives are good for operation sub-zero.

Still means trench digging of course, but theres no way round that without recourse to batteries and a big solar panel on the shed roof.

Shame no one in the UK manufactures motorised run off roof sheds that can be run from a PC. Theres a market niche.

Perhaps if my DIY To Do List was <1 sheet of A4 (8 point font).  :) 

Thanks again. Have you published any accounts of your observatory?

Offline
Last seen: 1 month 1 week ago
Joined: 05/03/2014 - 00:00
Observatory accounts

I haven't published that much Grant, the little that I did put online in various forums I removed.

Unfortunately I received the unwelcome attention of burglars. Not too much was taken from my older observatory, mainly damage caused by forced entry. The tool shed was ransacked and a mini lathe, mill and various power tools taken plus the house window frames damaged, though a combination of metal window frames and strong locks prevented entry.

The insurance company did cover the losses but I was asked about how much information existed about by hobbies and possesions in the virtual world and was warned that placing images or details of personal belongings on an open public forum was considered careless behaviour and that doing so would a) raise my insurance premiums by a hefty sum and, b) may lead to non-payment in future claims. I also had to extend the house alarm and CCTV to the observatory and outbuildings as a condition for continuing insurance protection.

Subsequently I have been more guarded.

Partly too, it has been a thirty year slow evolution to reach this point with nothing that damatic that I felt noteworthy at the time though looking back, moving from a concrete pad for a tripod to a fixed pier in the garden, through a wooden shed built to look like a summer house, followed by Skyshed Pod and now the Pulsar dome were big steps. Each change took many months and even years to complete with big gaps in continuity while work took precedence over hobbies and of course when I began the journey there were no on-line worlds to fill with a virtual life.

One day I will sit down and document the various stages my observatory has evolved through, though how to do that without giving away too many traceable personal details escapes me at present and first I must finish that home made spectrometer that I have been working on for the last eight years, then there is an ultra slim optical path Arduino-controlled rotator/focuser that I have been working on for a few years, still awaiting machining of the metal components to commence, I’ve had the aluminium billets since 2015 but somehow being retired has not given me as much ‘free‘ time that I thought it would....!

William.

Offline
Last seen: 3 days 10 hours ago
Joined: 11/03/2014 - 23:17
Hi William,

Hi William,

I note in your thread that you have a weather moniter, which model do you have and does it automatically close the dome shutter when rain is detected.

Cheers

Nick

Offline
Last seen: 1 month 1 week ago
Joined: 05/03/2014 - 00:00
Weather Monitor

Hi Nick

The principle weather monitor used is an AAG Cloudwatcher from Lunatico Astronomia. This has two (semi) independent methods of rain monitoring for observatory shutter control.

Normally, AAG software running on the observatory computer monitors the AAG hardware mounted on a pole adjacent the observatory where an I.R. sensor measures sky temperature to determine cloud coverage, a heated capacitive plate senses rain droplets directly and a light sensor detects sky darkness. I also installed the optional anemometer.

If user configured thresholds in the software for the I.R. sensor monitoring sky temperature are exceeded, water droplets fall on the capacitive plate, the sky brightens or high wind speeds are detected then the AAG software will issue a weather unsafe warning and the observatory management software (ACP) will initiate an observatory shutdown and issue a close shutter command to the dome controller. 

The above actions depend on the correct functioning of the observatory computer which can never be 100% reliable, even with UPS backup, so the AAG Cloudwatcher also has an independent pair of potential-free relay contacts on a dedicated socket built into the hardware mounted on the pole. When the capacitive plate senses water droplets the relay contacts close and via a separate direct wire feed back the the dome controller, the dome controller will close the shutter independently of any computer command.

The dome controller itself, which is a Rigel System from Pulsar Observatories (UK), uses a ‘heartbeat‘ monitor. If for any reason the observatory computer ceases to respond to regular ‘Are You Still Alive?’ queries then the dome controller will close the shutter. Also, as the dome controller communicates with the battery powered shutter via wireless it also has a ‘heartbeat‘ monitor and should the wireless link go down between shutter and controller then the shutter will independently close after a short delay.

The capacitive water sensor plate in the AAG Cloudwatcher hardware is heated to prevent condensation, melt ice or snow and dry the plate when rain stops. Because of this, power consumption of the AAG Cloudwatcher is quite high so the AAG hardware is powered from the mains supply via an AC/DC convertor.

Due to local topography and coastal location my observatory is subject to gusting wind speeds exceeding 90 mph several times a year. Should the open dome shutter be facing towards the wind it is possible the dome roof would be lifted off as the Pulsar Observatories dome design does not use a captive suspension system for dome roof to dome wall coupling, the dome roof rests on support rollers with a short skirt providing lateral location. As the observatory is under full automatic control of ACP I installed the optional AAG wind speed anemometer with the rest of the AAG pole mounted hardware. ACP monitors the AAG Cloudwatcher wind speed sensor and either will not startup the observatory and open the shutter or will close the shutter and suspend operations should windspeed exceeed my user defined threshold, currently this is set at a rather low windspeed of approx 30mph.

In practice, the AAG Cloudwatcher is not infallible as far as the I.R. cloud sensor works. The user defined thresholds need constant tweaking as the seasons progress and upper atmosphere temperature conditions vary, otherwise clear sky conditions are reported cloudy and vice versa. I had been working on writing a program in C++ code that could read my local public and ATC airport weather stations to automatically update the AAG thresholds for the I.R. cloud sensor but have found that my nearest airports ATC weather reports usually stop when the airport closes for the night so this project has been shelved for now.

My AAG unit is from an early batch and I read that Lunatico Astronomia have changed the specification of the I.R. sensor on the current shipped units to improve the stability of the the I.R. sensors response to upper atmosphere temperature changes.

The AAG Cloudwatcher also now includes a humidity sensor though apart from recording and reporting purposes I’m not sure it serves much purpose in active observatory management unless you incorporate this into dew heater management for example.

The AAG capacitive plate water sensor, light sensor and I.R. sensor are reliable but do need regular cleaning to keep them free of bird droppings etc so location for access is important.

At the time I purchased the AAG Cloudwatcher I also bought the AAG Solo weather server with the intention of mounting the AAG hardware high on one of the house chimney stacks, communication between the observatory computer and the AAG hardware would have been via the Solo unit and the LAN network. In the end I abandoned this approach for several reasons, partly planning since the house is a listed building but mainly accessibility to keep the AAG hardware clean (I did start down the path of building a heated washer system based around a salvaged automobile high pressure headlamp washer), possible issues with heat plumes from the chimmneys causing misreading of the I.R. sensor and the inabilty to use the built-in direct connection rain sensor relay contacts. In the end I decided to keep things simple and mothballed the Solo unit for now and placed the AAG Cloudwatcher hardware on a pole, 2mtr above ground and 1mtr from the observatory wall where it can easily be cleaned and wired directly to the observatory computer.

The anemometer is not heated and has frozen twice this winter where snow has melted and refrozen around the base. In practice as my location is just a mile inland of the South coast this has not caused a problem since we rarely experience snow, freezing conditions and high winds simultaneously but I will add an external heater cup to the base of the anemometer at some time. As an engineer I feel I can do this easily and at a much lower cost than Lunatico Astronomia’s own optional anemometer with built-in heater.

The ambient light sensor will respond to moonlight or domestic light spill so threshold settings need carefull adjustment to avoid false readings and unnecessary observatory unsafe conditions shutdown or delayed opening. Orientation of the AAG hardware so that the I.R. sensor faces the prevailing weather direction is important though there is no reason why you can not mount the AAG hardware on a pole that can be rotated to face into the prevailing weather direction and adapt to seasonal weather patterns.

Looking back at the observatory weather records for the last year I can see that with an observing session under way the AAG Cloudwatcher I.R. sensor detected cloud and initated observatory shutdown and shutter closure before it began raining on 37 nights while the capacitive rain sensor detected water droplets even though no cloud was detected and initated shutdown on 8 nights.

I did preferentially consider the Boltwood cloud monitor from Diffraction Ltd over the AAG Cloudwatcher The stability of the sky temperature detection of the Boltwoods active thermopile compared to the passive I.R. sensor used in the AAG Cloudwatcher is an advantage plus the no-moving-parts anemometer of the Boltwood is a simpler and more reliable device but I felt the higher initial purchase price and long term running costs of the Boltwood are too excessive for a purely amateur observatory. With a finite life of the thermopile and degradation of the thermopile window by airborne salt and organic contaminants, given my coastal  location and seagull - pigeon populations I would expect to have to return the Boltwood to Canada every three or four years for thermopile / window replacement.

William.

Offline
Last seen: 3 days 10 hours ago
Joined: 11/03/2014 - 23:17
Cloud Monitors

Hi Willaim,

I too have a Pulsar dome with a different moniter viz  HitecAstro Deluxe Weather Moniter, This requires a 12 volt supply and a serial connection to ther computer- I have a serial to USB converter; You set parameters relevent to your location and set e-mail and mobile phone alerts. In similar vein there is an independant feed to the Pulsar Control Box modified with an aditional socket allowing weather monitoring without a computer. Maybe AAG Cloudwatcher would have been a better choice as I have had a few issues in setting up the system. I am hovering over ACP as  I curently have an AP900 mount with APCC software SKY X and Maxim DL6 and set up each night with one target. Weather here on the South Coast has been appalling this winter with very few clear nights. CCD Navigator is also on my maybee list

Clear skies

Nick

Offline
Last seen: 1 month 1 week ago
Joined: 05/03/2014 - 00:00
Independent rain sensor

Hi Nick.

If you want an independent low power 12v rain monitor with switched output for the dome controller take a look at the Hydreon RG-11. I have one as a belt-and-braces backup for the AAG. The RG-11‘s switched output is wired in parallel to the AAG so either unit will cause Pulsar’s Rigel controller to close the shutter.

The RG-11 uses IR to detect rain droplets on a transparent dome, same as automatic windscreen wiper control in cars. It does have an internal heater to help evaporate condensation and works in a different way to a capacitive or resistive water sensor by reacting to sudden changes in reflected IR on the clear dome to trigger the relay so even if static water droplets are on the dome as long as no further droplets are seen then it will open the relay much earlier than the AAG unit.

http://www.weatherstations.co.uk/hydreon.htm

My experience so far is that cloud detection is more important than direct water droplet monitoring to prevent the optical equipment from rain damage since the sensing area of capacitive/resistive/optical rain sensors is rather small, the telescope can get quite wet before a single droplet is detected by rain detectors, I consider these to be fall-back safety backup devices rather than a primary controller.

I have mixed feelings about ACP as the observatory control software and was using CCD Commander for a while. Having a Paramount I am tied to having the TheSkyX running and ACP won’t work directly with TheSkyX’s automation interface. ACP requires Maxim and Focusmax so I have four separate applications running that often require updates to adapt to changes to Windows OS.

The only function in ACP that made it attractive was the scheduler engine. With our UK weather, being able to ask ACP to image a target automatically and keep acquiring data at the best times, however many nights / weeks / months that might need until all requested data had been captured, was the ‘killer’ application.

Software Bisque have been dropping hints that they are developing a new version of Orchestrate for TheSkyX that would offer many of ACP’s fuctions for the last four years or so but so far no news of progress or a release date and I couldn’t wait any longer so am committed to ACP, Maxim and Focusmax for some time to come. Certainly keeps you poor this hobby!

Regards.

William.

Offline
Last seen: 3 days 10 hours ago
Joined: 11/03/2014 - 23:17
CCD Autopilot

Hi William,

How would you describe the relative merits of ACP and CCD Autopilot?

Cheers

Nick

Offline
Last seen: 1 month 1 week ago
Joined: 05/03/2014 - 00:00
ACP v CCD Commander

Hi Nick.

I can't comment on CCD Autopilot, I did consider it and sent a couple of pre-purchase technical questions to CCD Ware that weren't answered so I gave it a miss. It still has a busy forum though so I guess there must be quite a few users out there.

I was using CCD Commander when I began automation and I found that very easy to use. The scripting interface is simple to understand and doesn't require a degree in computer languages to operate. CCD Commander worked very reliably for a managed observatory operation. By managed I mean remotely operated or locally operated but manually administered on a daily basis. Once a target or sequence of targets was programmed CCD Commander could be left to acquire the data and close up the observatory at the end of a session or shut down and restart a session in the case of a weather event.

I decided to move to ACP and ACP Scheduler mainly because of our fickle UK weather and by the limited unobstructed view afforded by my observatory location. I can only see around 55% of the hemisphere because of surrounding buildings and trees.

I found with CCD Commander I would start a target sequence and then we would have four or five weeks of bad weather, by the time the weather was clear the target was no longer accessible and after a few years I ended up with a large number of folders of incomplete targets and I found it hard to remember which target folders needed completion on any given night as the year turned and a particular target presented itself' again in favourable conditions.

The basic ACP program has not that much on offer to differentiate itself from other automation programs. The price is higher than others and on technical ability alone the high price is not justified but the programs author considers ACP to be a service and not a product. If you run into problems during the initial software support period, or if you buy a subscription for ongoing support you will be given one-to-one help, either by remote desktop direct to your PC, Skype, or regular telephone and a dedicated user forum. In this respect ACP can be considered a package, software and support, and providing this level of one-to-one support is obviously expensive hence the initial high purchase cost and annual subscription fees for ongoing support and updates.

If you have a fairly straight forward automation setup and can resolve most technical issues yourself, or are happy with a slower customer support then the basic ACP is hard to recommend on cost against technical ability, and bear in mind that ACP requires Maxim DL and FocusMax to function so the total software spend can be staggeringly large for an amateur observatory to carry.

For me, the power of ACP is only revealed once you buy the add-on, ACP Scheduler. Scheduler will look at all the incomplete targets you have begun as well as new targets and decide which target is in the most favourable position for imaging and will start acquiring data as soon as conditions are right. If weather stops the acquisition before completion and a few hours later the weather improves then scheduler will look at the current best target from the outstanding list and continue on with that instead of returning to the earlier one which may no longer be ideally placed. And of course, you can override this sequencing at anytime and go for a target of opportunity, follow up on a supernova alert, whatever, then go back to the original target sequence. All the while Scheduler keeps track of what needs doing and manages the observatory.

Since the move to ACP and ACP Scheduler two years ago I would say my acquired data and completed target rate has increased around 200% to 300%.

Initial setting up of ACP is complex and a first time user will need to make extensive use of the help files provided with the program, they are well written though and I was able to set up ACP and start running without needing support, I had more problems with TheSkYX and my Paramount MX than with ACP.

I don't know if the above is of any help....it's certainly a complex subject area and finding an automation product to suit your own needs can be difficult with so many competing applications on the market. Do make use of the free trial periods that most of the software vendors offer before jumping in to one package and be sure to determine what functions are essential for your automation project and which are wish-list items, it may help to narrow down the choice.

William.

Peter Carson's picture
Offline
Last seen: 3 days 6 hours ago
Joined: 06/04/2014 - 20:41
USB over Ethernet

Hi,

I've never used USB over Ethernet but I do know some who've tried but run into problems.
My observatory equipment is hooked up to a local cheap PC in the observatory and can be controlled via Real VNC by any device on the home wired network or even using my smart phone via the house wifi. Real VNC will enable control over the internet and I have been known to start the observatory going and then monitor it using my phone when I've had to go out in the car for a short while.

My first observatory PC lasted 7 years and the replacement PC's power supply blew up after a year but has been going for about two years since. The PC just has a plastic bag over it but no other precautions except that the observatory is regularly dehumidified.

Peter

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Real VNC

Have never heard of Real VNC - its 25years since I networked anything. Lots to read. Thanks.

callump's picture
Offline
Last seen: 1 day 5 hours ago
Joined: 03/01/2014 - 12:25
USB over IP

hey Grant,

I remember reading some articles about using a Raspberry PI to serve it's usb ports via Ethernet. At the PC end they just look like standard usb devices.

Some googling might turn up something helpful.

I did find this commercial offering https://www.virtualhere.com/

one device is free so might be worth an experiment.

Cheers, Callum

admin_dcf's picture
Offline
Last seen: 2 weeks 1 day ago
Joined: 10/01/2013 - 12:42
Linux vs Windows

My experience of weird USB configurations has tended to be that Linux seems a lot more tolerant than Windows. I'm guessing the kind of people who play with Raspberry Pis are much more likely to play with USB-over-ethernet, USB repeater-extension cables, etc, than the average Windows user. So I'm not sure it's fair to assume that just because this hardware works under Linux, the same will be true of Windows! :-)

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
USB over IP

So the Pi handles the USB control locally (and presumably provides buffering) and then squirts it across the network as loading permits? Thats sounds fun and worth trying out. Four Pis are hardly expensive - I recall that in the earlier models Pis were internally a single USB port so I would probably need at least 3 if I use two imagers.

Thanks.

jack martin's picture
Offline
Last seen: 4 hours 46 min ago
Joined: 20/01/2017 - 12:41
Grant,

Grant,

I use Team viewer or Tight VNC PC to PC to control the mount remotely via a 25m Maplin Cat6 Ethernet cable.

BEWARE you will need to protect / sleeve it when installing against foxes, it was bitten into but survived.

I can open / close the motorised roll off roof remotely with a key fob.

Regards,

Jack

Essex UK  

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Teamviewer, Tight VNC, Remote Assistance + others.

A key fob on the roll off roof is pretty cool :)

Will indeed look out for well protected cables. We have foxes round here certainly - the footprints in the recent snow were really obvious. Other things too - though pheasants, deer and voles are not normally a problem for telescopes, but rather unwelcome in the nearby veg patch.

I noticed several people mentioned the software they use to access the instruments in the observatory from indoors. Sounds like there are several good options out there - and some of them free to private non-commercial users. Its nice to have multiple good choices. I'm rather favouring using Dell D630 laptop outside as they have  4x real USB sockets and are rather cheaper than decent spec NUCs - t he NUC may come later (its compactness is attractive). Dell D630s are certainly not  the fastest machines in the world but good enough to run an autoguider and CCD at the same time.

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Update

I did try a Dell D630 30ft away from the house in the garden, with a CCD attached and it seemed to work fine when accessed by Remote Desktop and a Wifi connection. It happily sat there taking 5s exposures for me and displaying them. I didnt even need to move my Hub from the other end of the house - I will probably add a repeater when I use it in anger (theres minor issue of putting a  pier in, adding the mount/scope and adding an observatory to be overcome yet) or add a couple of ethernet cables to the trunking carrying the mains out there. 

I'm not sure why more people are not using Remote Desktop. What dumb error am I making? Security?  :)

Robin Leadbeater's picture
Offline
Last seen: 4 hours 26 min ago
Joined: 05/03/2014 - 00:50
I have run this way for over

I have run this way for over 10 years using cheap second hand computers running XP Pro in the observatory. They die after a few years in the less than ideal environment but I just replace them for little money. (I installed the 3rd one last summer). Recently I have taken to running it on a separate network to the house internet so I can keep the speed up and guarantee security with XP. (It is quite heavily loaded with guidescope camera (1s frame rate), guider (1s frame rate) and spectrograph imaging camera, mount, focuser and spectrograph controls)

Robin

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Networks

Separate networks sounds a good idea. Sneaky :)

nickjames's picture
Offline
Last seen: 4 hours 11 min ago
Joined: 12/10/2013 - 18:16
Indeed. I've been doing this

Indeed. I've been doing this for years with a cheap Dell in the observatory. My PC death rate is around one death every 2-3 years but I keep a disk image so replacement is dead easy. Windows RDP is certainly secure enough to use on an internal network. I wouldn't expose it directly to the internet though unless your network firewall can restrict access to particular IPs.

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Not so stupid

Glad to hear I havent got it completely wrong then. It is very reassuring to hear other people already do this.

I made sure the Dell I am thinking of is W7 rather than XP on because XP doesnt get security upgrades anymore. 

Now, if I can just get the garden sorted and the redecoration done and the skirting boards replaced and the paths laid and the.....

This will be fun.

Paul A Brierley's picture
Offline
Last seen: 4 days 20 hours ago
Joined: 05/03/2014 - 19:46
USB over Ethernet

Good morning Grant,

My CCD and DSLR-Mount-Guide camera are all controlled via USB through the one cable.

I have a Startech 4 port, powered USB hub which all the cables on that side of things are plugged into, on my Losmandy G11. I then have a Mutech USB Repeater Extension cable leading from this into the Whirlpool control-room. This cable is gold plated. I have the cable plugging into another Startech Powered hub.

I have never had any issues.

HTH.

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Sharing problems

Thats interesting. As my previous problem had been using a USB sound sampler on a shared port, I thought I should check again.

So, I was out last night using a USB GPS dongle and NMEATIME2 to set my system clock to <0.005s accuracy, controlling a telescope from TheSkyX  (which uses a  USB to Serial converter to talk to a Paramount)  and a Starlight camera via a USB socket.

The laptop I used was a Dell E4310 (Intel i5 at 2.4GHz with 4GB memory) which only has 2 USB sockets.

I noticed that when anything was sharing a socket with the dongle, the NMEATIME2 correction timings became erratic - it assumes GPS  fixes from the receiver are evenly spaced I think. So, when I connected the telescope or when sharing with the CCD my timing precision was potentially off as the attached image demonstrates (the sharp change in the trace is when I connected the telescope). Not surprising I suppose, but the timesharing aspect clearly does have some an effect on other bus users. I imagine the impact will depend upon the nature of whats sharing the port and what laptop manufacturer/model you choose.

I will try again tonight if (wonder of wonders) we get two clear nights in a row using a Dell D630 which has 4 USB ports and post the results here. Alas, I only have Dell laptops and so cannot say how this pans out with other machines. 

Attachments: 
Offline
Last seen: 1 month 1 week ago
Joined: 05/03/2014 - 00:00
Dell Latitude E4310 and shortage of USB ports

Hi Grant.

One of the issues you will face with laptops (and most desktop PC's) is that although you might have 3 or 4 physical UBS ports, on the outside of the laptop, inside the laptop all the ports are connected to a single internal USB root-hub which has an absolute limit on the data throughput from all the connected devices. I find that TheSkyX and my Paramount MX keep the USB root hub pretty busy as TheSkyX is continually polling the Paramount for Protrack drive corrections.

One way around this is that your Dell Latitude E4310 has an ExpressCard 34 slot, just above the DVD drive. If you plug in a ExpressCard 34 USB extension card, you gain direct access to the PC main bus without having to go through the existing laptop's USB Root Hub.

ExpressCard 34's are dying out now and becoming harder to find but some good quality brands still remain in stock with a few retailers.

They were available with two or four additional USB2/USB3 ports but recently, amongst the better brands, I notice only two port cards are still available and only USB3 cards are still commonly stocked, two port USB2 cards are getting hard to find.

The ExpressCard carries it's own internal root hub. Some cards also have the facility to connect an external power supply for the USB devices which can be handy for power hungry devices such as some guide cameras.

This gives you the option to put the low demand devices such as the GPS dongle, filter wheel, focuser etc on the laptop USB internal root-hub and your high demand devices like the Paramount and guide camera on an ExpressCard, or just put the GPS dongle on the ExpressCard alone. There are several ways to connect the devices about the laptops' and ExpressCards' available slots to keep the GPS dongle away from high demand devices.

Sample link to a current Startech 2 port USB 3 ExpressCard34:

https://www.ebuyer.com/338306-startech-2-port-usb3-0-expresscard-ecusb3s22

And a sample link to a current Startech 2 port USB ExpressCard34:

https://www.ebuyer.com/138357-startech-2-port-usb2-0-laptop-express-card...

The other option is to look for a second hand docking station for your laptop, that is if they ever made one, if you can find one that would usually give you another two/four USB 2 ports, again, directly connected to the laptops main bus and bypassing the laptops internal USB root-hub.

HTH

William.

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Thanks

Yes. That makes sense. Had wondered about getting an additional USB socket on an Express card. Hadnt thought of the docking station though (I do have one somewhere). Might get a spare docking station while they are still available - I have a feeling the later Dell models also have a connector, but its a different format - will worry about that when I upgrade (I'm a fan of kit thats oldish, so that when it breaks it is not an expensive event - good for the stress levels.).

Will grab a card and let you know how it turns out.

Grant Privett's picture
Offline
Last seen: 5 hours 42 min ago
Joined: 28/12/2014 - 18:30
Update

Just a quick update. As you may recall, I found that using a USB on my Dell E4310 meant the speed of the other USB socket was impacted. In fact, when a GPS dongle was being used to set system time, it meant the error could jump from 2-3 ms to 20ms.

To try to overcome this, it was suggested I try an Expresscard based USB port. I bought a Startech dual USB2 port card, but had to return it as it wasnt recognised on Win7 - even after 30 minutes with one of their helpful online advisors. So I tried instead a CSL dual USB3 socket Expresscard for £10. After loading the drivers that was immediately recognised. It  works  fine  with all the USB kit I tried. So I connected the GPS dongle up to that, hung a Paramount MEII tracking a satellite (22Hz tracking speed updates) on one of the laptop USBs and a Starlight 694 taking very short exposures on the other. Result can be seen above. For long periods the error was 2-3ms with occasional wandering to 4ms. Really not bad at all - plugging into the main bus made a big difference. Well worth the small installation effort.

So, if timing is crucial to you, get a standalone network enabled time source and set up your own NTP server. But if 2-3ms accuracy is good enough, £20 for a GPS dongle and a CSL card will do the business.