Build Your Own Davis Console Datalogger

[DeKay]

More good news.

Way back when I posted how to interface to the Davis console without having to buy Weatherlink and the dongle that came with it.  The limitation of that solution was that you didn't get data logger capabilities.  There was no way to get data from the console unless a computer was hooked up to it at all times.

Until now.

We have wxtech to thank for being crazy enough to unpot his logger and reveal what lurked beneath.  Since then I've been itching to see if I could build a DIY version, but my time has been very short.  That changed this weekend, and I have been able to successfully get something working.  More details can be found in my blog, specifically this post.

You can put one together for around $5 - $10 depending on what you have at hand.  I'd say that just about anyone who knows which end of a soldering iron is hot could put one together.  To add a bit of suspense to the whole thing, the memory chip I am using has four times the capacity of the Davis version (4 MBit vs 1 MBit).  I don't know if the console will actually make use of that extra memory or not, but I was willing to pay the extra 43 cents to find out.  I'll know in a few days and will let everyone else know.

Good luck should you try to build one yourself.  I'll keep an eye on this thread and the comments on my blog should anybody need a hand.

Here's a teaser.  And no, I do not own The Real McCoy. 

[C5250]

To add a bit of suspense to the whole thing, the memory chip I am using has four times the capacity of the Davis version (4 MBit vs 1 MBit).  I don't know if the console will actually make use of that extra memory or not, but I was willing to pay the extra 43 cents to find out.  I'll know in a few days and will let everyone else know.

It won't.

[graculus]

Once upon a time there was a company called BlueWave who made a clone logger, that had double the standard memory 

[Bushman]

Once upon a time there was a company called BlueWave who made a clone logger, that had double the standard memory 

But there is NOTHING stopping an individual from making their own.  And there is some doubt whether they can be commerically sold or not (Magnusson-Moss Act etc.)

[graculus]

But there is NOTHING stopping an individual from making their own. 

I completely agree, but how does that relate to the console's ability to support more than the standard logger memory?

[Bushman]

Hopefully Dekay will let us know shortly whether his logger supports double the mem.

[belfryboy]

Dekay, would you agree that this schematic agrees with your connections? I'm using a 3.3V FTDI cable attached to the 6 pin header.


VP_logger_schematic by belfryboy, on Flickr

And here is a layout (done in less than 1/2 an hour!) of the above. I have the design files and can make these for anyone and ship at cost plus a little for my time... (say about £40)


VP_logger by belfryboy, on Flickr

[m77]

wow.

youd need to buy a console separately, cos you'd need one to work with your logger, right?

@  £192 

[belfryboy]

can you buy a console separately, cos you'd need one to work with your logger, right?

Did that a few weeks ago, converted my wireless ISS to cabled (more reliable, and I can ) now just doing the logger thing, Thanks DeKay

[DeKay]

To add a bit of suspense to the whole thing, the memory chip I am using has four times the capacity of the Davis version (4 MBit vs 1 MBit).  I don't know if the console will actually make use of that extra memory or not, but I was willing to pay the extra 43 cents to find out.  I'll know in a few days and will let everyone else know.

It won't.

Sadly, C2520 is correct.  I topped out at 2560 records, which is the max capacity of the Davis product.  The BlueWave guys must have pulled an interesting trick to get more capacity.  Anybody know if they might have hacked the firmware to do so?  That would explain Davis coming down on them.  Oh well, the gamble was worth 43 cents.

[Weather Display]

Bluewave might have claimed the case, but not sure if it was really the case

[C5250]

I'll bet they didn't do anything except use a bigger PROM and leave it to ones imagination that it would store more records.

[DeKay]

Dekay, would you agree that this schematic agrees with your connections? I'm using a 3.3V FTDI cable attached to the 6 pin header.

<snip>

And here is a layout (done in less than 1/2 an hour!) of the above. I have the design files and can make these for anyone and ship at cost plus a little for my time... (say about £40)

Nice!  And fast!  Looks good to me.

Would you consider providing the design files under some kind of open license like the Creative Commons Attribution Share-Alike license so that people could spin their own boards?  That would very much be in keeping with the spirit of this whole thing, and nothing would stop you from providing a service to people for ready-built units.  Think Arduino.

[C5250]

No offense intended... Looks like the raw output of an autorouter. I would have at least tweaked it to eliminate the trace going through vias.

[DeKay]

No offense intended... Looks like the raw output of an autorouter. I would have at least tweaked it to eliminate the trace going through vias.

Electrically it is correct, but the layout could indeed use work.

The trace from pin 1 on the SOIC should be re-routed so it doesn't go through pins 2 and 3.  That would be more DIY friendly by reducing the risk of solder bridges there.  The two vias on either side of BLK could be eliminated completely.  Vias are annoying for anyone doing a homebrew PCB.

[Weather Display]

checking on some very old threads, it could be that it did actually hold more data
so I take that back

[Grrxyn]

No offense intended... Looks like the raw output of an autorouter. I would have at least tweaked it to eliminate the trace going through vias.

Electrically it is correct, but the layout could indeed use work.

The trace from pin 1 on the SOIC should be re-routed so it doesn't go through pins 2 and 3.  That would be more DIY friendly by reducing the risk of solder bridges there.  The two vias on either side of BLK could be eliminated completely.  Vias are annoying for anyone doing a homebrew PCB.

Not to continue shooting a dead horse, but I would have taken an extra step and added a bypass capacitor for the flash chip.

To add a bit of suspense to the whole thing, the memory chip I am using has four times the capacity of the Davis version (4 MBit vs 1 MBit).  I don't know if the console will actually make use of that extra memory or not, but I was willing to pay the extra 43 cents to find out.  I'll know in a few days and will let everyone else know.

It won't.

Sadly, C2520 is correct.  I topped out at 2560 records, which is the max capacity of the Davis product.  The BlueWave guys must have pulled an interesting trick to get more capacity.  Anybody know if they might have hacked the firmware to do so?  That would explain Davis coming down on them.  Oh well, the gamble was worth 43 cents.

Any more results on the IM-ME device?  Would be neat to strap a radio interface to an Arduino or similar and have a home-brew console.

[belfryboy]

No offense intended... Looks like the raw output of an autorouter. I would have at least tweaked it to eliminate the trace going through vias.

not autorouter, I wouldn't touch the stuff, I'd agree with you about the routing, but I was on my lunch hour with a sandwich in one hand. I would be happy to share the design files with anyone that would like them. Likewise I would be happy to assemble them too.

Electrically it is correct, but the layout could indeed use work.
The trace from pin 1 on the SOIC should be re-routed so it doesn't go through pins 2 and 3.  That would be more DIY friendly by reducing the risk of solder bridges there.  The two vias on either side of BLK could be eliminated completely.  Vias are annoying for anyone doing a homebrew PCB.

I will try and sort out the layout today to make it easier. I suppose that i'm lucky that I have a board house that will make me free prototypes

[belfryboy]

here is the updated layout, I may even change the 6 pin header to SMD so that all the board can be made single sided.


VP_logger_mkII by belfryboy, on Flickr

thanks for all your input guys!!

[DeKay]

Any more results on the IM-ME device?  Would be neat to strap a radio interface to an Arduino or similar and have a home-brew console.

Nope.  Progress postponed due to summer.  But summer is over and my time is freeing up to get back in to this stuff.  It is still somewhere on my list of things to do.

[DeKay]

here is the updated layout, I may even change the 6 pin header to SMD so that all the baord can be made single sided.

While you are at it, you could add an FTDI chip and a USB connector to the schematic and layout.  These could be stuffing options so the board could accommodate both a FTDI cable and a regular mini-USB interface.  An example schematic and layout from the GoodFet open source project is here if you'd like to take a look.

[belfryboy]

While you are at it, you could add an FTDI chip and a USB connector to the schematic and layout.  These could be stuffing options so the board could accommodate both a FTDI cable and a regular mini-USB interface.  An example schematic and layout from the GoodFet open source project is here if you'd like to take a look.

I already have this done!!  I simply used the sparkfun design files and gave them a tweak. there's even room to put a legacy serial interface on the board.

I do prefee the FTDI cable option, since these are readily available and reduce risk for the DIYer.

[C5250]

here is the updated layout, I may even change the 6 pin header to SMD so that all the baord can be made single sided.

That looks better! I always cringe when I look at a board and see unneeded vias.

I, briefly, thought about one sided. I don't think there's as much cost advantage these days as there used to be though. Unless you're going to make the board without silkscreen and solder mask.

[belfryboy]

here is the updated layout, I may even change the 6 pin header to SMD so that all the board can be made single sided.

That looks better! I always cringe when I look at a board and see unneeded vias.

I, briefly, thought about one sided. I don't think there's as much cost advantage these days as there used to be though. Unless you're going to make the board without silkscreen and solder mask.

I'd agree but if your etching yourself then single sided is easier, if your getting it made by a board house then vias make it easier to lay out. Out of interest do you lay out many of your own boards? I'm fairly new to this and trying to learn, but always appreciate constructive criticism.

[DeKay]

I'd agree but if your etching yourself then single sided is easier, if your getting it made by a board house then vias make it easier to lay out. Out of interest do you lay out many of your own boards? I'm fairly new to this and trying to learn, but always appreciate constructive criticism.

I know this was directed at C2520, but I'm hoping to get setup to etch my own boards sometime this winter.  I am debating UV vs. toner transfer vs. direct resist application via inkjet printer.  Either way, single sided certainly makes life easier, both for the board itself and any kind of solder reflow scheme like toaster ovens or hot plates that might be used to actually assemble it.

Thinking of a home process, you could easily be less aggressive on your ground pour and increase the trace spacing a bit to make the design easier to DIY.  There is much more process error at home vs a commercial board house where shorts could creep in.