Those of you who have read any of my posts will know that I have had disconnection problems with my Davis VP2 USB data logger. I have tried various patches etc. but the problem persists. Trawling through posts on this site and others is making me think that the only true solution is to replace it with a serial logger ( not sure if Davis, or my supplier here in UK would now exchange it after almost a year ) . Conspiracy theories abound as to what causes the disconnect everything from bad USB leads supplied by Davis to using a paper shredding machine in the same room , switching a lamp on etc. etc. Personally I’m sure that it disconnected on at least 2 occasions when one of my kids signed off from MSN on the same computer but maybe I’m just getting paranoid and sucked into this conspiracy stuff ( was that really Elvis that I saw in our local fish and chip shop ? ) Apparantely Davis have tried to recreate the disconnect but report that they find no problems. I saw that Brian ( Windy ) said on the " Weathermatrix " forum that serial is 100% successful, so I guess my question is if I do buy a serial logger will I definitely overcome this problem ??? Also I think I read somewhere that you cannot use a USB to serial converter on a laptop ( obviously no chance of fitting serial port ) ; can’t understand why that would be ??
Cheers,
Many of those USB users who do experience disconnect problems nowadays find that one of:
** Using the v5 USB driver;
** Using a USB lead with ferrite cores attached at each end;
** Ensuring that there are no sources of interference nearby or feeding interference into the same electrical ring;
cures the problem. But there are still a few installations for which nothing seems to work and maybe yours falls into this category?
There should be no problem that I’m aware of about using a serial-to-USB adapter with a laptop - many users of the serial logger do this (but provided it’s a compatible adapter - most are, including of course the 8434 Davis-branded one, but there are some, especially with older USB chipsets that aren’t.)
Don’t forget that the 6555 IP interface logger is due to be released shortly (‘late 2007’). This will be more costly than the standard one but also more flexible though full details aren’t yet available and maybe the standard serial logger will do all that you need.
Firstly I would avoid the V5 driver - Davis have pulled it off their web site. In my system it caused fatal errors and blue screens at intermittent times. A real pain as I was blaming WD 10.37f which I loaded at the same time. Version 4.4 is the one to go with. The ferrite cores and avoiding interference are sensible options along with disabling selective suspend mode on the USB hubs (My Computer>properties>hardware>device manager>universal serial bus controllers>USB root hub>properties>power management>‘off’). Also change bps to 9600 in the ports setting (my computer>properties>hardware>device manager>ports>CP210x USB to Uart>properties>port settings>9600). Finally try running Virtual VP which will at least keep everything running and give you an audit of how often, if at all, the logger is dropping data.
Like you I am based in the UK and the Serial Data logger is x2 the price compared to the US (clearly someone does not understand the concept of ‘exchange rates’ - $136 =
I will try your suggestions next time I am in France ( remote station and not yet got remote software installed ). I would love to get the USB logger to be stable rather than change to Serial. I’ll let you know.
Unfortunately many of us trusted our UK supplier who made no comment on potential USB problems! But yours are wise words for those about to embark on a purchase.
I must admit to being fortunate in having procured a serial data logger at the outset, having noted discussions on the various wx forums. However, it does seem odd that with so much spoken about the USB conundrum (connectivity problems) that Davis still continue to push the USB logger. This, when the serial logger would appear to be faultless with a “good” Serial/USB convertor for those who do not have a serial port on their PC. Perhaps this is because the Davis USB Datalogger driver is outsourced and there are legal implications. But it would seem that some of the Retail sources should be made more fully aware and not be almost in denial that a possible problem exists with USB connectivity.
There has been mention on this forum previously, that when challenged by USA customers, Davis has readily changed USB data loggers for the Serial type when requested to do so. This would almost appear to be an admission that in reality the USB datalogger is “not fit for purpose” in all circumstances. Perhaps the UK importers (McMurdo) and Davis have a better part to play with the UK Retail dealers and their customers.
Yes, I’m aware that there is going to be an IP version soon, but in the meantime…???
I think I need to say this again because it can be in danger of getting lost in these discussions about the USB logger: The great majority (say 90-95%) of customers buying the USB logger find that it operates to their satisfaction with no problems. But certainly yes there is a small minority who do experience the disconnect problem and are understandably quite vocal on the various weather forums (leading to misconceptions about how prevalent the problem actually is). And even amongst this minority, one or other of the fixes mentioned previously can often be successful. I would say over the past year for example that the USB loggers that we’ve needed to swap out for the serial version (which we’ve done quite willingly) because no other fix worked have been down at the 1-2% level of units sold.
I don’t want to go over the background to this problem in detail again here because it’s been written about extensively in previous threads here and elsewhere, but (i) a reputable dealer will indeed typically talk about this issue before taking an order - I know we often take some pains to do so (though sometimes a customer is still adamant that they want the USB version or thinks we’re trying to make an extra margin by offering the 6510SER + 8434 combination), so choose your dealer with care; (ii) this always has been an elusive problem when it comes to trying to recreate the problem and pinning down the cause, very possibly because there is more than one cause and context. I know that Davis have expended considerable effort in trying to do so and that various solutions like changing the USB chipset have appeared promising but then have, on extended testing, failed to eradicate the problem completely.
As to fitness for purpose, I think one of the problems with the great diversity of PC hardware and software configurations and operating circumstances eg re local RFI - this diversity/flexibility of course being the prime driver of the success of the generic PC design - is that virtually no PC accessory or ancillary item can be sold with guaranteed compatibility; there are always combinations of circumstances in which a particular item fails to operate successfully. The USB WL logger is no different from many other devices in this.
If anyone in the UK has experienced a problem trying to swap out a USB logger when they’ve had genuine disconnect problems, then it’s time to consider whether they’ve used a reputable dealer. The swap-out scheme is fully operative in the UK for genuine UK-sourced parts, has been for quite a while, and there shouldn’t be a problem taking advantage of this. It’s obviously better to get the swapout arranged as soon as the problem is spotted and not many months or years later (and of course if it’s not a UK-sourced part then you’ll need to go back to the supplying dealer.)
Sorry I meant to comment on this point too. To the best of my knowledge, v5 has not been available from the Davis site as yet - new SiLabs drivers are never available for download from Davis until they’ve been extensively tested at Davis and that process is still ongoing for the v5 version - there may be some more information on the results in a week or so. Davis did very recently reorganise the software download section of their site and maybe that left a misleading impression of what was available before and after.
It’s noteworthy that you’ve had a bad experience with the v5 driver although yours is the only adverse report that I’ve seen so far. Some other users have the v5 driver loaded without problems and seeing some benefits for the USB disconnect problem. I’ll see if I can remember to provide an update as soon as I have some further information.
And while adding to previous posts, I should have included in the list of possible fixes the use of VVP, which again some users have found helpful (with the USB reconnect option enabled). This is part of the problem here I’m afraid, there are several possible fixes all of which seem to work for some particular PS installations/configurations but not for everyone and there still are a few folk left for whom unfortunately nothing seems to work.
still, to avoid being one of those 5 % of people, simply get a serial data logger and a USB to serial convertor (good quality one).
I honestly dont know why the hardware retailers dont promote that more
Sadly that’s not an infallible combination either - I’ve certainly known the 6510SER WL serial logger plus the 8434 USB adapter combination to suffer from the USB disconnect problem also, but yes it is of lower incidence than 6510USB with the original USB driver. My take on this is that for the standard PC architecture, running a data logger continuously 24/7/365 is a relatively demanding task and shows up wrinkles that are simply not evident in shorter-term use. In this environment although USB is by and large reasonably robust (thought sadly this is not the experience for every single user) it still doesn’t quite match up to RS232 serial all-the-way for reliability.
I guess it’s partly because it makes the data logger that many people already consider to be very expensive even more expensive. And serial->USB converters don’t always work well. I bought another one identical to the one I use for my serial VP2 data logger to use on another PC that is too modern to have real serial ports. The converter clashes with another (expensive) bit of USB kit (not weather related) and stops it working. The supplier of the expensive kit thinks the problem is in their code and is working on a fix, but for now I have to choose between using the expensive kit or the various serial related devices I have. Unfortunately the expensive kit software can interface with some other serial devices and I can’t use that functionality at the moment
but every person who had trouble with the usb data logger and changed to a serial data logger and then purchased a good quality USB to serial data logger has had no more problems, that I know of, that I have recommended it to
I have to wonder how much of this problem is due to Davis putting a usb/serial chip into the existing design rather than putting the effort into designing a “real” usb interface logger :?
That’s good and long may it continue! I agree - there’s no doubt that 6510SER+8434 (for example) is likely to be a solution with a good chance of success for that limited percentage of users who have encountered problems with 6510USB plus an older USB driver. But don’t be too surprised if you do start to encounter cases where it’s not a complete solution because I’ve certainly seen examples myself. That’s partly what I’m trying to say: 6510USB with an older USB driver may be a more vulnerable combination, but to a degree I’d argue that it’s the USB protocol in general that’s not 100% robust to 24/7/365 operation. If people ask, my preferred solution is RS232 all the way for maximum reliability, but yes if they must use a USB input, for example on a laptop, then 6510SER+8434 is probably the next best thing in the reliability stakes.
I’m not sure there is such a thing as a ‘real’ USB solution. I’ve not looked at USB chipsets extensively, but a couple I’ve seen are really just protocol converters between serial data streams and USB protocol. A good example is the chipset (FTDI) used in my serial->USB converters. The chips in the converter are the same chip as used in another piece of USB equipment I have that has a data transfer rate in the megabits/second range.
Oh, maybe I misunderstood :oops: I had the impression that other USB data devices, eg the LabJack, use the HID (like mouse/kb) standard and I assumed that was not the same protocol/chipsets as used by usb/serial devices.
…and where I was heading is that in years of having USB mouse/trackballs/touch pads that have sat idle for a long time I have never had the experience of one losing contact with the mothership or not responding when touched, so to my non-expert impression IF that’s a different interface/protocol it would appear to be more robust than the one the USB weatherlink is using.
If you look at ‘Windows Supplied Drivers’ on this page url]http://www.sss-mag.com/usbp2.html[/url] there are some limitations of HID devices (low bandwidth, limited packet sizes, etc). Those limitations aren’t an issue for keyboards/mice, but may well be for other types of devices. It may also be the case that HID programming is more difficult than serial port type programming and in the case of the VP2 data logger Davis have to code for a serial device so making the USB device use as much of the same code as possible makes the programming job much easier.
I’d guess that this is a key point. Although it dominates the market in reasonable-spec but affordable weather stations, Davis is not a large company in the big scheme of things and has relatively limited development resources, especially for software. AIUI a significant chunk of that resource is currently being used for final testing & development of the IP logger interface and driver. It may be that, in the long run, this IP interface is seen as a standard interface for the WL logger, ie possibly rather than developing a ‘native’ USB interface, but that’s pure speculation.
But if the IP interface were to prove to work reliably in long-term use with no significant problems and the IP interface could be added at minimal cost then, especially in the light of past USB inteface problems, it might be tempting for Davis to promote that as the standard. There are obviously some issues to think through eg would it always be a cabled IP connection in the interests of maximum reliability (in which case you would need to work out how the cable would conect to a PC whose network socket was already in use for a LAN or broadband connection, do you assume everyone would need at least a minimal hub/switch etc) or might there say be a Davis WiFi adapter. I’m speculating wildly here but you can see the sort of possibilities that start to open up once you have a proven IP interface.)
