OR... could it be that the story although true was created as an allegory... to Feynman the fruit loops had disappeared from the Universe, but in reality we know that they must be somewhere. And we know that to find them you would have to know approximately where they might be. Some of those places might be pretty sensitive and you might require permission.
After all that discussion about RS485 cable I decided to wing it with some crappy 18 gauge 4 core. It's that brown cable you can see a few inches of in the photo above. The contractors got it strung up yesterday. A day later than expected.
All was working well. Data was coming down the pipe. I left it running over night and it was still OK in the morning. Then I spent the day fixing up some software configuration. When I got back to the hardware all was dead. No data, nothing, nada.
After much poking around two things became clear:
1) The power supply for this, that I had brought over from Finland and has been working well for a year or more, was now shutting down for whatever reason.
2) The RS485 to USB adapter, which has also been in use for a year or so, is now dead.
The power supply thing is obvious, my Fluke shows zero volts after it has been running for a while.
As far as we can tell there is very nice RS-485 signal coming down the pipe. When fed with a good PSU, an expensive looking Agilent in this case. Good voltage levels, good clean edges. No weird noise spikes. But no data coming out of that USB adapter anymore.
WTF? Is it possible I have two random hardware failures at the same time? Or is there a common denominator?
Back to Fry's. At least they have some decent 24v power supply modules. Not a 485 adapter in sight though.
Looks like I could be stuck here for another week.
Moral of the story:
When your boss says fly a million miles and get the job done with zero days notice and no site survey and nothing prepared, just say no!
I have to say, the young guy in Fry's took it as a challenge to find an RS485 to USB adapter. Searched the shelves, check their computer stock, sadly failed.
1) The power supply for this, that I had brought over from Finland and has been working well for a year or more, was now shutting down for whatever reason.
2) The RS485 to USB adapter, which has also been in use for a year or so, is now dead.
WTF? Is it possible I have two random hardware failures at the same time? Or is there a common denominator?
Smells like something common. How is your ESD pathway handling ?
Isolated RS485 starts to look appealing...
Speaking of Fry's, after what seemed like an eternity cruising the 100Km of isles, I was happy to see Parallax in stock there. Could not resist picking up some bits and bobs. Got me a ColorPal, no idea what I would do with it but seems like a cool thing to have.
What cable shield? We ain't got no namby pamby shield. This is 400ft of 18 gauge, solid copper, four core cable. Not even twisted pairs. This is real mans cable, you could drive an electric trolly bus with this stuff.
I'm now stuffing 24 volts in and our device will run down to 7 volts so resistive losses are no problem. We took a look at the signal coming down the pipe with a scope on Friday evening. Looks good, nice clean edges, nice voltage levels, it's only 11,5200 baud, no sign of any noise getting in so that should be fine.
Thanks for the USOPTL4 suggestion. That looks perfect. Isolated every which way, LEDs to see what's happening, terminal block connection. Great. I've asked the boss to order a bunch of them.
Does anyone know how quickly Mouser delivers ?
@jmg,
Yes, I can't so easily believe these two failures, PSU and adapter, happening at the same time are random independent events. Something killed them. Now that distant memories come to mind I have seen earthing problems kill RS485 interfaces before. In one case I was called to investigate the grounds at each end were flying up and down by 60 volts at mains frequency. And both units were in the same cabinet connected to the same mains supply! Instant death when they were connected together. Sorting the ground fixed that.
@Phil,
Yes, earth the shield at one end. Something I have heard a lot, never had occasion to worry about. Luckily we don't have that problem, there is no shield! Shields are for wimps
Thanks. That's good to know. That is what we will do.
To be honest I'm not sure about the common ground. The contractors have bolted a pipe to the wall and mounted the device just above roof level. Any grounding happening there is purely by accident. I'm not even sure if the aluminium chassis of that thing is connected to the internal circuitry. I checked a long while ago and now can't remember. I'll have to buzz out another one we have in the office in the morning.
What does it do?
I'm going to leave that as a mystery for now. Some of you may be able to guess from the picture.
If I ever get it working properly all will be revealed. If not, I'm keeping shtum and trying my best to forget all about it.
That might be an incentive for every one to help out when I get stuck again
It's not even the main point of this project, which is more about software, annoying hardware detail.
I just got the green light for staying in San Jose for another week.
There is now time to sort this mess out. I wish I had been given more than three days notice to prepare. Still that's a lesson for the boss. Or perhaps for me to say "no way".
I wonder what problems will hit us next....
I was hoping to get home a mid-summer trip to our lake side cottage 500Km into the forest. It's a kind of national party time. On the other hand it will probably snow then even if we did have a super hot spring time.
I'm now stuffing 24 volts in and our device will run down to 7 volts so resistive losses are no problem.
What is the end of line voltage ?
Maybe that bump to 24V has just stressed enough the Power supply ?
What margin does that have over 24v ? - could it tolerate 48v spikes ?
I would check carefully for GND effects, as those are quite high voltage looking lines nearby, best to isolate the far end from ground, and use an isolated 485 bridge.
I have no idea what voltage is at the end of the line. To find out I'd have to climb up to the roof via that ladder. See earlier picture. No way, I'm amazed I made it up there the first time!
My back of the envelope calculations, and online voltage loss calculators, tell me there should be a 6 or 7 volt drop down the line for the current this thing consumes. It can run from 7 to 32 volts so 24 should be good to go.
The 24 volts I have now comes from a huge big and expensive Agilent lab power supply. The old supply went in the trash. I'm hoping that thing has no issue. I'm going to be run out of town if I blow that up!
That's a good point about the high voltage lines running down the street. A farmer neighbor of mine once ran an electric cow fence wire down the middle of a field under some high voltage lines. There was enough voltage induced in that wire to keep the cows away even without a "ticker" running.
I don't see an mains frequency on the receiving end but I guess spikes can happen.
Certainly we will get hold of an isolated 485 interface.
Manage to extend your trip out to June 22nd and you'll be able to attend the San Jose Sensors Expo/Conference. The boss should support that! Might as well fit in that trip to see Chip too.
About Mouser, if you order more than $250 worth of stuff, one perk you can pretty much count on is a shipping upgrade from UPS ground to 2nd day air. Phone in your order to be sure. A couple of those nice USOPTL4's plus a bit more would get you there.
Heater, you are barely high enough on that roof to get away from a barking dog plus the Fire department is right next door to come and rescue you! I know the unit is facing directly north off the top of the ReStore building but that is also the controversial San Jose Environmental Innovation Center and Prospect Silicon Valley plus Household Hazardous Waste.
So this unit might be green tech, and could have something to do with the environment, like tracking the sun and measuring UV/IR and pollution etc?????
Ha, on the global scale of things that roof is indistinguishable from ground level. But there is no way I'm going up that ladder again. It's not getting stuck that bothers me, it's heart failure and dropping like a stone. Besides I'm not insured for such things.
The ReStore place is fascinating. I've never seen anything moving in or around the Household Hazardous Waste building.
No, we are nothing to do with "green" and the environment. At least not directly. In fact apart from the windmills outside that almost never turn and are only for show anyway, and the solar panels over the parking lot, I don't see anyone there doing anything related to green/environment exactly.
So, now, what is the controversy surrounding San Jose Environmental Innovation Center/Prospect Silicon Valley?
Prospect is a neat "office hotel" for your little start up I think. There is office space, workshop space, a bit of a machine shop. Think of it as Obama's Hacker Space
That was a harder sell than I thought it would be. The boss claims he was at that Expo two years ago and that there was nothing interesting there. Still, he is not a techy.
Regarding your previous power supply issue, the mains voltage is 120VAC/60Hz in the US, and if it wasn't auto-ranging then there's often a 110/230 switch on it. Contrary to how a device with a switch set for 110V plugged into 230V systems instantly blows up, a 230V device plugged into 120V might actually work- barely and intermittently.
Embarassingly, once I was involved with a machine sent to a 50Hz region with a fluorescent tube light fixture, and even with properly re-tapping the transformer to get 120V AC from the local power system the light barely lit, flickered, and buzzed.-- apparently US 120V fluorescent tube ballasts are so cheaply made that they only work at exactly 60 Hz.
Ah yes, I thought I had that covered. The supply in question claims it can handle 100/250 volt AC input. No mention of frequency though, as far as I recall. It's in the trash now.
However, you may well have a point, this is the first time it was used outside Europe.
I got hold of a new supply to replace it, can't use that big Agilent lab supply forever, it does indeed have a mains voltage select switch.
I was amused, years ago, to see the guys in our lab operating some US gadget from a huge audio amplifier fed with a 60Hz sine wave. Turned out the thing used mains frequency for timing so that was their solution.
Comments
Works for me.
But what are four of erco's larger robots doing in a cheap van doing watching all of this at work?
After all that discussion about RS485 cable I decided to wing it with some crappy 18 gauge 4 core. It's that brown cable you can see a few inches of in the photo above. The contractors got it strung up yesterday. A day later than expected.
All was working well. Data was coming down the pipe. I left it running over night and it was still OK in the morning. Then I spent the day fixing up some software configuration. When I got back to the hardware all was dead. No data, nothing, nada.
After much poking around two things became clear:
1) The power supply for this, that I had brought over from Finland and has been working well for a year or more, was now shutting down for whatever reason.
2) The RS485 to USB adapter, which has also been in use for a year or so, is now dead.
The power supply thing is obvious, my Fluke shows zero volts after it has been running for a while.
As far as we can tell there is very nice RS-485 signal coming down the pipe. When fed with a good PSU, an expensive looking Agilent in this case. Good voltage levels, good clean edges. No weird noise spikes. But no data coming out of that USB adapter anymore.
WTF? Is it possible I have two random hardware failures at the same time? Or is there a common denominator?
Back to Fry's. At least they have some decent 24v power supply modules. Not a 485 adapter in sight though.
Looks like I could be stuck here for another week.
Moral of the story:
When your boss says fly a million miles and get the job done with zero days notice and no site survey and nothing prepared, just say no!
I have to say, the young guy in Fry's took it as a challenge to find an RS485 to USB adapter. Searched the shelves, check their computer stock, sadly failed.
Fry's by night (Why aren't I down the pub?):
Isolated RS485 starts to look appealing...
-Phil
I hope the boss does not check the bills to closely I sneaked a few goodies in to the company purchases.
Check Fry's vs the Mayan's :
@evanh,
What cable shield? We ain't got no namby pamby shield. This is 400ft of 18 gauge, solid copper, four core cable. Not even twisted pairs. This is real mans cable, you could drive an electric trolly bus with this stuff.
I'm now stuffing 24 volts in and our device will run down to 7 volts so resistive losses are no problem. We took a look at the signal coming down the pipe with a scope on Friday evening. Looks good, nice clean edges, nice voltage levels, it's only 11,5200 baud, no sign of any noise getting in so that should be fine.
Thanks for the USOPTL4 suggestion. That looks perfect. Isolated every which way, LEDs to see what's happening, terminal block connection. Great. I've asked the boss to order a bunch of them.
Does anyone know how quickly Mouser delivers ?
@jmg,
Yes, I can't so easily believe these two failures, PSU and adapter, happening at the same time are random independent events. Something killed them. Now that distant memories come to mind I have seen earthing problems kill RS485 interfaces before. In one case I was called to investigate the grounds at each end were flying up and down by 60 volts at mains frequency. And both units were in the same cabinet connected to the same mains supply! Instant death when they were connected together. Sorting the ground fixed that.
@Phil,
Yes, earth the shield at one end. Something I have heard a lot, never had occasion to worry about. Luckily we don't have that problem, there is no shield! Shields are for wimps
I presume that both ends have a common ground?
And just what does the gadget on the roof do?
Thanks. That's good to know. That is what we will do.
To be honest I'm not sure about the common ground. The contractors have bolted a pipe to the wall and mounted the device just above roof level. Any grounding happening there is purely by accident. I'm not even sure if the aluminium chassis of that thing is connected to the internal circuitry. I checked a long while ago and now can't remember. I'll have to buzz out another one we have in the office in the morning.
What does it do?
I'm going to leave that as a mystery for now. Some of you may be able to guess from the picture.
If I ever get it working properly all will be revealed. If not, I'm keeping shtum and trying my best to forget all about it.
That might be an incentive for every one to help out when I get stuck again
It's not even the main point of this project, which is more about software, annoying hardware detail.
I just got the green light for staying in San Jose for another week.
There is now time to sort this mess out. I wish I had been given more than three days notice to prepare. Still that's a lesson for the boss. Or perhaps for me to say "no way".
I wonder what problems will hit us next....
I was hoping to get home a mid-summer trip to our lake side cottage 500Km into the forest. It's a kind of national party time. On the other hand it will probably snow then even if we did have a super hot spring time.
Like it did last year. See here:
Maybe that bump to 24V has just stressed enough the Power supply ?
What margin does that have over 24v ? - could it tolerate 48v spikes ?
I would check carefully for GND effects, as those are quite high voltage looking lines nearby, best to isolate the far end from ground, and use an isolated 485 bridge.
My back of the envelope calculations, and online voltage loss calculators, tell me there should be a 6 or 7 volt drop down the line for the current this thing consumes. It can run from 7 to 32 volts so 24 should be good to go.
The 24 volts I have now comes from a huge big and expensive Agilent lab power supply. The old supply went in the trash. I'm hoping that thing has no issue. I'm going to be run out of town if I blow that up!
That's a good point about the high voltage lines running down the street. A farmer neighbor of mine once ran an electric cow fence wire down the middle of a field under some high voltage lines. There was enough voltage induced in that wire to keep the cows away even without a "ticker" running.
I don't see an mains frequency on the receiving end but I guess spikes can happen.
Certainly we will get hold of an isolated 485 interface.
About Mouser, if you order more than $250 worth of stuff, one perk you can pretty much count on is a shipping upgrade from UPS ground to 2nd day air. Phone in your order to be sure. A couple of those nice USOPTL4's plus a bit more would get you there.
I love you! What a cunning an devious idea! That expo sounds fabulous, it's going need at least two days, right?
As it stands, next week I loose a day on Tuesday anyway, the site will be over run my visitors as they put on some other demos there.
And I need a day to got to ITS America in San Jose.
I'm really glad I did not call the boss earlier and agree to the extra stay already, he would have bought a ticket for the 20th or something.
He should be amenable to the sensors idea. He keeps hassling me about such things.
Heck I'll ask for a flight on the 23rd. After all this has already messed up my holiday back home, right?
Do they sell Propellers at Mousers...
So this unit might be green tech, and could have something to do with the environment, like tracking the sun and measuring UV/IR and pollution etc?????
Ha, on the global scale of things that roof is indistinguishable from ground level. But there is no way I'm going up that ladder again. It's not getting stuck that bothers me, it's heart failure and dropping like a stone. Besides I'm not insured for such things.
The ReStore place is fascinating. I've never seen anything moving in or around the Household Hazardous Waste building.
No, we are nothing to do with "green" and the environment. At least not directly. In fact apart from the windmills outside that almost never turn and are only for show anyway, and the solar panels over the parking lot, I don't see anyone there doing anything related to green/environment exactly.
So, now, what is the controversy surrounding San Jose Environmental Innovation Center/Prospect Silicon Valley?
Yay, I'm in. Flight home on the 23rd June.
That was a harder sell than I thought it would be. The boss claims he was at that Expo two years ago and that there was nothing interesting there. Still, he is not a techy.
Regarding your previous power supply issue, the mains voltage is 120VAC/60Hz in the US, and if it wasn't auto-ranging then there's often a 110/230 switch on it. Contrary to how a device with a switch set for 110V plugged into 230V systems instantly blows up, a 230V device plugged into 120V might actually work- barely and intermittently.
Embarassingly, once I was involved with a machine sent to a 50Hz region with a fluorescent tube light fixture, and even with properly re-tapping the transformer to get 120V AC from the local power system the light barely lit, flickered, and buzzed.-- apparently US 120V fluorescent tube ballasts are so cheaply made that they only work at exactly 60 Hz.
Ah yes, I thought I had that covered. The supply in question claims it can handle 100/250 volt AC input. No mention of frequency though, as far as I recall. It's in the trash now.
However, you may well have a point, this is the first time it was used outside Europe.
I got hold of a new supply to replace it, can't use that big Agilent lab supply forever, it does indeed have a mains voltage select switch.
I was amused, years ago, to see the guys in our lab operating some US gadget from a huge audio amplifier fed with a 60Hz sine wave. Turned out the thing used mains frequency for timing so that was their solution.
Unexpectedly, I am by no means the oldest person in here!
Let's see...
400' of healthy mains cable.
Heater on the roof surveying his domain.
One dead power supply.
It must be a Propeller powered portable electric Bessemer converter, readied to pour molten iron over intruders as they approach your castle.
Failing that, an rf receiver of sorts, not a transceiver as you're only interested in data coming down the pipe.
Although configuration data will need to be sent up the pipe sometimes.