What's the Real Story Behind USB Over Cat 5/6 Extenders?

[Scorpius]

OK - so I’m thoroughly confused regarding how USB over Cat 5/6 extenders actually work. All the manufacturer info I’ve read indicates the transmitter is located at the PC and the receiver is located near the mount, cameras, etc. The usual wording is the transmitter “sends” a signal to the peripherals but what about signals that need to travel in the other direction – like from the camera to PC when the images are downloaded? In other words, do these USB over Cat extenders carry a signal in both directions as this would seem absolutely necessary to return what the camera has captured to the PC?

My observatory is currently under construction but I need to start making some educated decisions regarding data transfer so any insight would be greatly appreciated. My desire is to have hard wired connections for everything as opposed to separate PC’s networked together and wireless solutions.

In my situation, the distances involved aren’t huge but definitely exceed the standard 5m for USB and in light of some connectivity issues I’ve already experienced, I’m after a dependable solution even if it costs a little more than I’d like.

This is the 4-port USB over Cat 5/6 device I’m considering but as usual - the manufacturer talks about sending the signal to remote devices but says nothing about receiving any data back. Surely, I’m just making the whole data transfer process more complicated than it actually is but could someone please humor me by explaining how these USB/Cat devices actually function? Also, has anyone experienced issues running multiple devices – like an imaging camera & guide cam - through a single Cat 5 or 6 cable connected to a multi-port hub?

B203104PNP.pdf

[Macavity]

As you may have deduced USB extenders can cost "anything" - from a "few quid" to very much more.

For a discussion of the USB standard, probably better to get your information closer to source?

https://en.wikipedia.org/wiki/USB

You may "get lucky" with low cost devices. Some work "mostly", but can lose contact at levels sufficient

to drive you to distraction eventually! Give that experience, I gave up on the budget units and invested

in one of the ICRON Rover devices. These include full data buffering and provide handshake signals

to simulate standard 5m cables. Personally, I only need USB 1.1 speed to control a mount + focusser

etc. I have a "industrial strength" metal case 4-port Hub at the remote end to distribute signals...

For higher (camera) data rates devices for USB 2.0 / 3.0  (including hubs) over Cat5 / fibre optic exist.

Since you are located in the USA, you are probably better served for distributors!

Culturally, the methodology seems more popular too! The Icron company page is:

http://www.icron.com/products/icron-brand/usb-extenders/

In the UK, most people seem to prefer to use a scheme based on having a Laptop

(temporarily / permanently) in the observatory and linking via TeamViewer (qv). 

But I operate in a fairly cool, damp, seaside location which seems to cause certain

problems with more delicate electronics and mechanics.

P.S. you may find the links don't work first time (spurious characters appended)?

[JamesF]

The signals must go in both directions, otherwise it wouldn't work.

You can extend USB to a certain extent by adding an "active" repeater.  This is basically an inline USB hub powered from the USB port itself.

The devices that use cat5 or fibre I'm really not sure about.  I assume they have some sort of USB hub inside and the data from that gets transmitted using some other protocol to the remote end where it is converted back and made available by another USB hub.  USB places strict requirements on latency though and if you're using these links for large amounts of data continuously (eg. a planetary camera) I can't help but wonder if the cracks might start to show.  I don't know that they will, but the opportunity appears to exist   Even more so if you wanted to do the same with USB3.

It may be that there are ways to deal with some of the latency issues by making the intermediate hardware much faster and/or more complex, which is perhaps what the really expensive models do (unless the business is run by Martin Shkreli).

James

[Scorpius]

As you may have deduced USB extenders can cost "anything" - from a "few quid" to very much more.

For a discussion of the USB standard, probably better to get your information closer to source?

https://en.wikipedia.org/wiki/USB

You may "get lucky" with low cost devices. Some work "mostly", but can lose contact at levels sufficient

to drive you to distraction eventually! Give that experience, I gave up on the budget units and invested

in one of the ICRON Rover devices. These include full data buffering and provide handshake signals

to simulate standard 5m cables. Personally, I only need USB 1.1 speed to control a mount + focusser

etc. I have a "industrial strength" metal case 4-port Hub at the remote end to distribute signals...

For higher (camera) data rates devices for USB 2.0 / 3.0  (including hubs) over Cat5 / fibre optic exist.

Since you are located in the USA, you are probably better served for distributors!

Culturally, the methodology seems more popular too! The Icron company page is:

http://www.icron.com/products/icron-brand/usb-extenders/

In the UK, most people seem to prefer to use a scheme based on having a Laptop

(temporarily / permanently) in the observatory and linking via TeamViewer (qv). 

But I operate in a fairly cool, damp, seaside location which seems to cause certain

problems with more delicate electronics and mechanics.

P.S. you may find the links don't work first time (spurious characters appended)?

Macavity - Thanks for the insight & links. The Icron Ranger series looks like the Cadillac of USB over Cat 5 extenders but the price certainly makes it a tough decision. If I decide to bite the bullet and go for the best technology, this is the model I’m considering: http://icronshop.com/home/usb-20-ranger-2204

The signals must go in both directions, otherwise it wouldn't work.

You can extend USB to a certain extent by adding an "active" repeater.  This is basically an inline USB hub powered from the USB port itself.

The devices that use cat5 or fibre I'm really not sure about.  I assume they have some sort of USB hub inside and the data from that gets transmitted using some other protocol to the remote end where it is converted back and made available by another USB hub.  USB places strict requirements on latency though and if you're using these links for large amounts of data continuously (eg. a planetary camera) I can't help but wonder if the cracks might start to show.  I don't know that they will, but the opportunity appears to exist   Even more so if you wanted to do the same with USB3.

It may be that there are ways to deal with some of the latency issues by making the intermediate hardware much faster and/or more complex, which is perhaps what the really expensive models do (unless the business is run by Martin Shkreli).

James

Hi James – I appreciate the info on powered USB hubs and the potential pitfalls of USB over Cat 5 extenders. I found this statement in Icron’s documentation which seems to indicate their Extreme USB technology addresses the latency issue you mentioned but who could know for sure unless proven by actual experience with their products?

“Put simply, the LEX manages the expectations of the host computer and the REX manages the expectations of the USB devices. In both cases, Extreme USB keeps the link active during the increased delays resulting from the link extension.”

[JamesF]

Yes, I'm not surprised that some people try to do that.  It's a logical thing to attempt.

What it probably can't address though is the expectations of the software on the computer and, for example, the firmware on a camera at the remote end.  If you have a camera that can be configured to run at 50fps say, what happens when, thanks to the increased signal transmission time, it can't actually push 50fps down the connection?  Some systems will restrict their transmission rates if they know that bandwidth is limited, but I don't think it's generally considered in driver code and suchlike that latency may be more than you'd expect to find based on the USB specs.

If I were buying one I think I'd reasonably expect it to work if the USB link were not being heavily used, but if I tried to use it for a system that ran a "normal" USB connection close to saturation and it didn't work then I'd not cry about it.

James

[Scorpius]

Yes, I'm not surprised that some people try to do that.  It's a logical thing to attempt.

What it probably can't address though is the expectations of the software on the computer and, for example, the firmware on a camera at the remote end.  If you have a camera that can be configured to run at 50fps say, what happens when, thanks to the increased signal transmission time, it can't actually push 50fps down the connection?  Some systems will restrict their transmission rates if they know that bandwidth is limited, but I don't think it's generally considered in driver code and suchlike that latency may be more than you'd expect to find based on the USB specs.

If I were buying one I think I'd reasonably expect it to work if the USB link were not being heavily used, but if I tried to use it for a system that ran a "normal" USB connection close to saturation and it didn't work then I'd not cry about it.

James

I realize I'd need to provide more specific equipment specs to expect a conclusive response but do you think it would be reasonable to expect the Icron extender I posted a link to above to handle an imaging camera plus a guide cam with the addition of a motorized filter wheel & focuser at some point in the future?

You're obvioulsy a more logical thinker than me but if I spent over $300 for a USB extender and it couldn't do the job - I'd probably cry like a big ole baby!!!

[jonathan]

I built a custom serial cable from about two feet of Cat5, not sure how relevant that is to this conversation other than to say it's just copper at the end of the day whether it's got moulded USB plugs on the end or a hand-soldered plug from Radio Shack.  I'm sure the latency thing is what would prevent the use of a long piece of wire for USB, but perhaps you could utilise some other tech such as the Raspberry Pi in place of that very expensive looking USB extender?  USB signal comes in, small program to send the data out of the network socket, receive at the other end maybe with a second Raspberry Pi or just as a raw data signal into the USB port, capture the data using some software.  Might involve some port programming but I have seen it done for RS232 serial data into a USB socket and it worked very well.

[JamesF]

I realize I'd need to provide more specific equipment specs to expect a conclusive response but do you think it would be reasonable to expect the Icron extender I posted a link to above to handle an imaging camera plus a guide cam with the addition of a motorized filter wheel & focuser at some point in the future?

You're obvioulsy a more logical thinker than me but if I spent over $300 for a USB extender and it couldn't do the job - I'd probably cry like a big ole baby!!!

The filter wheel and focuser won't generate much traffic and it will all be between exposures, so I don't see them as a major issue.  And in fact the guide cam is most important during exposures, when the main camera (I'm assuming we're talking about long exposure DSO imaging here) will not be generating traffic either.  So I think there's a reasonable balance there that makes it more likely than not to work.

But £300 is a lot of cash for that sort of thing.  You could easily get an acceptable laptop to run it all from locally for that and then just run an ethernet cable back to wherever you need it and control everything remotely using something like VNC.

How long a cable run do you need?

James

[JamesF]

I built a custom serial cable from about two feet of Cat5, not sure how relevant that is to this conversation other than to say it's just copper at the end of the day whether it's got moulded USB plugs on the end or a hand-soldered plug from Radio Shack.  I'm sure the latency thing is what would prevent the use of a long piece of wire for USB, but perhaps you could utilise some other tech such as the Raspberry Pi in place of that very expensive looking USB extender?  USB signal comes in, small program to send the data out of the network socket, receive at the other end maybe with a second Raspberry Pi or just as a raw data signal into the USB port, capture the data using some software.  Might involve some port programming but I have seen it done for RS232 serial data into a USB socket and it worked very well.

There is existing code to do this sort of thing, so it may not involve that much work as long as you have the right skills.  I've not tried it myself however, so I can't speak for the quality of the project from personal experience.

Experience with house electrics and such like can make it tempting to think that electrics is not that different from plumbing.  It's just pipework for electrons, right?  That may be an acceptable first approximation for mains or low voltage DC circuitry, but once you get into playing with high frequency digital signalling it isn't like that at all.  It ceases to be "just copper".  Signals can "bounce" off inadequately-terminated cable ends and cause interference with later signals.  The "skin effect" can become an issue.  Capacitance between neigbouring wires can cause problems.  RFI may be a problem.  Electronics isn't really my bag, but I am aware that things can generally get a bit weird.

Serial line signalling rates are fairly low really.  Lots of hardware runs at 57,600 baud or considerably less.  You can get away with quite a bit (and I often have   I don't think I've ever cabled something up to work faster than 115.2k though.  That's small potatoes compared with USB3, which from memory is up in the range of several Gigahertz.

James

[Scorpius]

The filter wheel and focuser won't generate much traffic and it will all be between exposures, so I don't see them as a major issue.  And in fact the guide cam is most important during exposures, when the main camera (I'm assuming we're talking about long exposure DSO imaging here) will not be generating traffic either.  So I think there's a reasonable balance there that makes it more likely than not to work.

But £300 is a lot of cash for that sort of thing.  You could easily get an acceptable laptop to run it all from locally for that and then just run an ethernet cable back to wherever you need it and control everything remotely using something like VNC.

How long a cable run do you need?

James

Well James for starters I’d like to say your depth of knowledge on a wide array of subjects never ceases to amaze me. I know your insight has helped me on several occasions as well as a lot of other folks here on SGL.

So here’s my situation the  best way I can describe it. Although I’m currently in the process of building a ROR observatory http://stargazerslounge.com/topic/247778-windy-knoll-observatory-my-build-thread/ ,I’m still using a pop-up tent as a makeshift control room. I just don’t like being joined at the hip to the scope and on most nights the dew here falls so heavy that anything not covered gets soaked, including myself. So I’ve been trying to push the 16 ft. USB limit (one 10 ft. & one 6 ft.) to the imaging cam & guide cam in order to get the tent far away enough from the scope that it doesn’t obstruct the view. But it’s not working out & I’ve had all sorts of issues trying to do it this way. Early on, I purchased one of the cheap, powered USB hubs but haven’t seen where it’s improved things at all. It’s not just the cabling either as I continue to have issues with various software crashing and it very well could be my Windows 8.1 laptop that’s the root of the problem. However, I’m at a loss trying to figure out if it’s the PC, software, cabling or a combination of all those things and the last time I set up the conditions were better than they’d been in months. But wouldn’t you know, I couldn’t get anything to work including PHD which refused to guide no matter what I did. It started throwing error messages about the Celestron ASCOM driver and it was so bad I finally got frustrated and packed it in with nothing to show for my efforts. Therefore, since I can’t re-write the software and drivers and I’m not ready to take a chance on a new Windows 10 PC, I decided to start with the cabling to see if a quality USB extender would solve my problems.

The total distance I currently need from the tent (on the ground next to the unfinished observatory) - through the conduit already installed under the floor - then up through the pier & out the owl’s nest to the cameras is roughly 30 ft. Once phase I of construction is complete (main observatory with control pod) that distance may drop slightly but then when I build an actual control room next year It will go back up to 30 ft. (or slightly more) since the cables will be routed through a different, longer  conduit which terminates inside the pier.

So what I was hoping to do is find a reliable USB solution that will work for my current situation all the way through until the observatory is complete. Maybe a 2’nd PC located in the underfloor compartment next to the pier would be the best approach but right now, trying to figure out what the least expensive, most reliable, hassle free way to proceed is driving me up a wall. As far as I can tell, it’s not the mount, OTA or cameras that cause most of the problems but for me, it’s the software, drivers, cabling & PC that have been my nemesis ever since I first got started. And I must not be the only one since I’m constantly seeing posts about PHD not guiding, compatibility issues with software and drivers which begs the question – with all the money we spend on gear - why can’t (or why won’t) the manufacturers offer solutions that just plain work out of the box? Why don’t I want a new windows PC? – because I’ve yet to see any manufacturer that says it’s drivers are compatible with that OS and if it won’t even run on Windows 8, why should I believe it will work on Windows 10?

If this has turned into a rant then my apologies but we shouldn’t have to spend even more money on stuff to get our equipment (which already costs a small fortune in the first place) to function. But if it takes a new PC or a $300 extender to get everything to consistently function as designed then my attitude at this point is - I don’t really have much choice unless I just give up which isn’t going to happen any time soon...

[beamer3.6m]

I like to keep it simple with my no frills obsy and I think cables and connections are one of the areas where gremlins can get into a system easily.

My solution was to keep a very cheap laptop (cost me £30 on gumtree) in the obsy. This runs the mount/focusser/dslr/guidecam with no issues.

The laptop is about 2 feet from the pier so short cables.

I then sit indoors with my cup of tea and a more powerful laptop and control it all via teamviewer.

It is so simple that it just works.

In the event that teamviewer goes down or the internet dissapears I can use the cheap laptop in the obsy direct. Ok it wont be any good for image processing but the reality is that the workhorse pc/laptop does not need a lot of grunt as other than the usb and a half decent amount of memory it sits and does very little.

I know this does not answer your question on cat5 cables but try not to het lost in the complication of it all.

Simple works.

[Scorpius]

I like to keep it simple with my no frills obsy and I think cables and connections are one of the areas where gremlins can get into a system easily.

My solution was to keep a very cheap laptop (cost me £30 on gumtree) in the obsy. This runs the mount/focusser/dslr/guidecam with no issues.

The laptop is about 2 feet from the pier so short cables.

I then sit indoors with my cup of tea and a more powerful laptop and control it all via teamviewer.

It is so simple that it just works.

In the event that teamviewer goes down or the internet dissapears I can use the cheap laptop in the obsy direct. Ok it wont be any good for image processing but the reality is that the workhorse pc/laptop does not need a lot of grunt as other than the usb and a half decent amount of memory it sits and does very little.

I know this does not answer your question on cat5 cables but try not to het lost in the complication of it all.

Simple works.

Beamer - thanks for the advice and description of how a second PC easily resolves the USB distance limitation and simplifies the entire process. Sounds like what's often referred to over here as the KISS method (keep it simple stupid) applies to amature astronomy as well...

Regards,

Scorpius

[Zakalwe]

The only problem with using a low powered control PC in the obsy is if you intend to do lunar/planetary/solar work. The modern cameras need SSDs and USB3 sub-systems.

I personally think that this is a very interesting solution:

http://www.cloudynights.com/topic/505276-myt-no-cables-no-laptop-no-problems/

Mounting a small form factor PC directly on the mount and then using Remote Desktop to control it. Then just a single power lead to the mount and a single power lead to the onboard PC/power distribution.

[Macavity]

SGL (as with way forums) has "it's way of doing things" in certain areas?

Based on direct experience, I occasionally post alternative/tested solutions.

I fairly (rashly) "tip the wink" to others who I KNOW have different opinions!

I rather understand the OP's frustration: "Best do it as it's always done"?

It would be unfair to suggest a hardware "non-starter" to any questioner.

But where is the direct evidence that fast USB links definitely don't work?

Can we not progress from [iMO] *awkward* (albeit working!) solutions?

Of course, the ability to experiment depends on one's budget. But frankly,

If I had accepted "it can't be done" (on your budget) advice, I'd have given

up long ago. Let's not automatically exclude new ideas. Hey, Just sayin!

[beamer3.6m]

The only problem with using a low powered control PC in the obsy is if you intend to do lunar/planetary/solar work. The modern cameras need SSDs and USB3 sub-systems.

Well I suppose it depends on your equipment but I stand by my post. Simple works. Yes the KISS principle.

So you need a laptop with usb 3. Thats pretty common and can still be picked up cheap.

My qhy5 works fine as does my dslr. Is that cutting edge equipment. No. Is it sufficient for me yes.

In reality I think usb2 is fine for 30fps imaging.

As for ssd. Buy one and install it. My £30 laptop has one installes by me. Simple.

The pc on a scope looks awesome but too complicated for me.

[JamesF]

I'm certainly not saying it won't work, just that there may be configurations when it won't, or at least won't be reliable.  If I were going to drop that kind of cash on something then I might want to have a chat with the vendor before hand and agree that if it didn't work they'd at least take it back and refund the cost.  Or even talk to the manufacturer and explain what is intended to check their opinion.

As I posted previously, with long-exposure DSO imaging the demands are likely to be quite different from lunar/solar/planetary imaging and I think there should be a reasonable expectation that it would work.

If I were in the position of the OP I think I personally would go back to basics and get one thing at a time working reliably with the PC close to the mount and telescope even if it meant reinstallng the PC from the ground up, then look at the options for more remote operation.  That's just the way I prefer to do things though.  I'm a very boring and methodical problem-solver as a rule

James

[Zakalwe]

Well I suppose it depends on your equipment but I stand by my post. Simple works. Yes the KISS principle.

So you need a laptop with usb 3. Thats pretty common and can still be picked up cheap.

My qhy5 works fine as does my dslr. Is that cutting edge equipment. No. Is it sufficient for me yes.

In reality I think usb2 is fine for 30fps imaging.

As for ssd. Buy one and install it. My £30 laptop has one installes by me. Simple.

Yes, USB2 is fine for 30fps but modern planetary cameras can run at hundreds of frames per second. Couple that with a large sensor and USB3/Gig Ethernet and SSDs become mandatory. I have tried running a ASI 174 camera on a 7200RPM hard drive and it constantly dropped frames. KISS also means using the right kit for the job as otherwise you end up bodging with workarounds.

An onboard PC is actually a simpler solution than USB over Ethernet or multiple USB hubs and active cables. All of the leads are kept to the minimum length. There's little or no chance of cable snags or cable drags. Cable routing becomes a non-issue. Probably the only thing is making sure that a simple cover is placed over it to prevent dew from soaking it.

[Macavity]

Interesting new idea, Z... A fair Ampere-age to think about, maybe?

Don't mind be being a grumpy git, folks? Not ALL to do with SGL lol

Personally, I am fairly happy going out there with my trusty Notebook

for lunar/planetary/solar imaging. But I still drive everything else via 30m

of USB 1.1 Cat 5. In general, problems have been bad connectors.

But then, all my VIDEO astronomy signals come via standard coax.

Old style "integrating" Video Cams had to have some thing going...

And, surprisingly, that was even part of the planning.

[andyo]

Hi if you don't mind paying the cash then you pick these up from anywhere between £250 -£400.I emailed Startech and asked them about an issue James raised in another post

"Another is that some USB device drivers expect the device they control to respond within a given timeframe, or make assumptions about the actual total travel time for the signals between the PC and the device."

Startech advised that while they can't test the exact setup I gave them they are confident that this device will work and not  delay the signal so the driver fails.

http://www.startech.com/uk/Cards-Adapters/USB-2/USB-Extenders/4-Port-USB-2-Extender-over-Cat5-or-Cat6-Up-to-330-ft-100m~USB2004EXT2

You can pick these up for anywhere between £250-£400 not cheap!!

I personally use a startech active repeater with a startech externally powered 7 port hub. Apparently its not a good idea to use an active repeater with a usb powered hub and you have to be careful when mixing an active repeater with an usb hub as there could be compatibility issues with the electrical components used in them.

[michael8554]

Hi Scorpius

Looking at the specs of the cameras you have, they are all USB 2.0, the fastest being the Orion SSAG Pro at up to 200fps.

Like many SGL readers I had multiple USB issues when I built my Obs and warm-room, which are 7 metres apart.

The system that finally worked was using individual Active USB 2.0 cables, as in my case I have an XP Tower PC with 8 USB 2.0 sockets, plus a proper RS232 port for scope control.

Not having a 200fps camera I found that cheap 7M Active USB 2.0 cables from the Bay work without problems.

So I see no reason why individual 10 dollar 10M Active USB 2.0 cables from the Bay shouldn't work for all your low-speed devices, and maybe for your SSAG too.

If not you should have enough change from 100 dollars to buy one quality branded cable for the SSAG.

If insufficient USB bus power to a device is causing problems, there are externally powered Active cables too.

Michael

[D4N]

Yes, I'm not surprised that some people try to do that. It's a logical thing to attempt.

What it probably can't address though is the expectations of the software on the computer and, for example, the firmware on a camera at the remote end. If you have a camera that can be configured to run at 50fps say, what happens when, thanks to the increased signal transmission time, it can't actually push 50fps down the connection? Some systems will restrict their transmission rates if they know that bandwidth is limited, but I don't think it's generally considered in driver code and suchlike that latency may be more than you'd expect to find based on the USB specs.

If I were buying one I think I'd reasonably expect it to work if the USB link were not being heavily used, but if I tried to use it for a system that ran a "normal" USB connection close to saturation and it didn't work then I'd not cry about it.

James

You might like to know that oacapture seems to handle this fine, I was using it to remotely test my asi120mm connected to my PI and forwarded over wifi. The frame rate was obviously right down but otherwise it worked fine

I think it is probably not a great approach for planetary but since I only plan to guide with it that's not a problem.

/Dan

Sent from my iPad using Tapatalk

[JamesF]

I have some plans regarding remote high frame rate cameras, but they're a bit of a way down the queue at the moment

James