Internet Steam Pressure Gauge

Before our Steam Rocket leaked fuel into the top of it, I used to have a wireless basestation on my desk, and I’d got in the habit of looking at the activity lights when a download was taking longer than it should to see if anything was actually happening. Being a creature of habit, I kept glancing at where it used to be whenever I was waiting for a web page; clearly a replacement was needed.

Well, here at last it is: a steam gauge to measure the pressure in the tubes of the Internet.
This is a real steam pressure gauge, but with a slight modifcation - it now measures packets per second. More pictures and construction details are here.

I hope the inappropriate mixing of technologies appeals to you too!

Inappropriate mixing of technologies is what we are all about!

Once again, you deliver originality and fantasticness in a tightly bundled package, I take my hat off to you Prof Fzz, over and over again

I have a brass pressure guage sitting right next to me...it is begging to be converted! If only I knew the secret language of the ones and zeros.

You even managed to keep the Bourdon tube and linkages! Good show, that man!..

Cheers
Harold

By Jove — that servomotor contraption is such a simple solution I doubt it would ever have occurred to me!

Rowan of Rin wrote:
I have a brass pressure guage sitting right next to me...it is begging to be converted! If only I knew the secret language of the ones and zeros.

I have recently ordered the Arduino Starter Kit from Maker Shed, which package contains an Arduino Diecimila microcontroller board, an assortment of components, and a copy of the O'Reilly/Make book Making Things Talk. Once it arrives (I expect it within the next week) and I've had time to play with it, I shall post a review.

My non geeky wife is in awe. I guess that means I need to find a pressure gauge and convert it.

Z.

Very nice.

Being a total programming klutz, any chance of seeing the code on your website someday?

Flonker wrote:Very nice.

Being a total programming klutz, any chance of seeing the code on your website someday?

Sure, though I really should add some comments first, because it will need tailoring for other uses - too many magic numbers in the code.

Excellent!!!

What a devillishly cunning chap you are.

Most definitely the best invention I've seen in a long time!
I am most impressed by the retention of the original mechanism rigged up to the new one rather than its complete replacement!

Pheobsky wrote:
I am most impressed by the retention of the original mechanism rigged up to the new one rather than its complete replacement!

As soon as I took the back off and realised how elegant the original mechanism was, I knew I couldn't just remove it. It probably would have been simpler to remove the old mechanism and just drive the pointer directly from the servo (though you only get about 120 degrees of motion from this servo). But it just didn't seem the right thing to do - the mechanism has done many years of service and didn't seem ready to be scapped quite yet.

The downside of using the old mechanism is you don't get such good precision - if you come to a value from above and then come to it from below, the needle doesn't settle on precisely the same value. My paperclip linkage isn't exactly a high-precision solution. But it's certainly good enough to get a rough idea of network traffic levels, and that's all I really wanted.

Stark-raving-GENIUS!

I'm in awe, old chap, AWE I tell you!

Flonker wrote:Very nice.

Being a total programming klutz, any chance of seeing the code on your website someday?

I've commented the code and linked to it from the relevant place in the text on my website. It should give you the general idea, but I think anyone building a similar system will have to tailor it somewhat (fix the relevant USB port ID, figure out the zero values, etc).

The first piece of code is servo-set.c which sets the servo position or speed. You need to read the web page for the SD21 to figure out the registers to set. For servo 1, for example, you set register 0 to set the speed and register 63 to set the position. You could use the 16-bit registers, but there's not much point given all the other inaccuracies. And try not to drive the servo into its stops, as this isn't good for them. However on the micro servo I have there are no stops, so specifying too small or too large a value causes it to spin endlessly at high speed - not good if you've got a meter attached.

The second piece of code is meter.tcl which monitors the network and calls servo-set to drive the meter. This is not a good example of coding, but it does the job. I wouldn't really recommend TCL as a scripting language these days - Python would be a better choice. But for other reasons I work with TCL a lot, so it's easy for me to script in it without thinking. This piece of code has far too many magic numbers in it, such as the servo position for zero on my gauge, the range needed for full scale deflection on my gauge, etc. Don't assume you can use it unchanged! But it might give some idea of how my version of the gauge works.

Oh, and all this code is for MacOS X. It should run on Linux or FreeBSD with very minor tweaks. I've no idea how you get network stats on Windows I'm afraid.

Well, I must admit this has aroused a significant degree of cupidity in me! I need to buy me a servo controller.

Oh, and one thing I forgot to mention - while servos aren't noisy, they're not completely silent either. This could get really annoying if it was in a bedroom. In my study though, the quiet noise of the gauge makes is quite reasuring, indicating that everything is working as expected. Though my Mac Mini is pretty silent, computers usually make all the wrong noises, not related to what they're actually doing as real machinery should do. This fixes that to some degree...

Utterly fabulous sir, and Mac OSX code to boot!

Now then, where can I get a suitable gauge and servo from...

Quite.

I just came across this single servo USB controller. If you just want to drive one servo, it's considerably simpler and cheaper than the boards I'm using. However, I have other plans for the spare 20 channels on my servo controller board...

Professor Fzz wrote:I just came across this single servo USB controller. If you just want to drive one servo, it's considerably simpler and cheaper than the boards I'm using. However, I have other plans for the spare 20 channels on my servo controller board...

You forgot to go "MWA-HA-HA!"