Ground Loops..
- eGoCeNTRoNiX
- Posts: 7362
- Joined: Wed Oct 23, 2002 12:51 pm
- Location: HELL
Ground Loops..
Ok Guys.. I've done a little bit of research on this subject of ground loops. We have a scale at my 2nd job that weight cotton. The scale sends output to the computer via wedge. We have had problems with the scale going beserk lately and can only attribute it to a possible ground loop. Because if the scale is not connected to the computer it is fine. But as soon as you hook it up to the computer it goes crazy! It has calmed down a little bit, but not enough to solve the problem. After weighing about 300 bales (they weigh 2000 a day) the scale will freeze up. Unplug it and plug it back in and it's fine. The only thing that we can think of is since the scale is grouned and then the actual head unit for the scale is also grounded and the computer is grounded that we are creating a ground loop which is causing the scale to go bonkers. Anybody know a way around this? Would cutting the 3rd prong off of the power cord going to the computer damage it? I know that it's not good to do it, but I've got to figure out something to keep this from happening as it slows down production. (1 hr costs them about $1k) I don't even know if cutting off that prong would fix the loop. But it would make sense to me? eGo
PM before Email People!!
Heat Under eGoCeNTRoNiX
Who Farted? BEANIE!!!
!Welcome to the United States of the Offended!
Heat Under eGoCeNTRoNiX
Who Farted? BEANIE!!!
!Welcome to the United States of the Offended!
- FlyingPenguin
- Flightless Bird
- Posts: 33162
- Joined: Wed Nov 22, 2000 11:13 am
- Location: Central Florida
- Contact:
I've had a LOT of experience with ground loops concerning video, sound and data. I worked for a company that provided computers, sound and video services at race tracks and ground loops are a curse with long wire runs.
Ground loops are caused when there's a difference in voltage potential between two locations. Electrical code allows electricians to uses the neutral instead of a ground in electrical boxes, assuming that the neutral is grounded at the breaker box. This can aggravate ground loops.
First off, you may NOT have a ground loop problem. You may have a data drop-out problem. Maybe the signal isn't reliably reaching the computer because it's too weak, or because of poor shielding, or some electrical motor or transformer that's not properly grounded is causing noise in the circuit.
Is this RS-232 serial data? How far are we talking?
Are you using a signal amplfier? RS-232 serial data is limited to a hundred feet without shielding if I remember.
What we used to do was use a device on each end that converted the RS-232 to another serial standard (RS-432? not sure) which CAN be sent very long distances on an unshielded twisted pair with no common ground. Regular phone wire works fine. We've run it as far as 5000 feet.
How far is the computer from the scale? Is there any way, even if only temporarily, to run them both off the same power circuit with a long extension? If it's possible to do this then you can confirm that the problem is indeed a ground loop. If the problem goes away when the scale and computer are on common power then it's almost certainly a ground loop, if not it's something else.
If the above test isn't possible then you'll have to experiment.
Yes you can "float" the ground on the computer and it might help (either cut off the ground prong or use a ground lifter). HOWEVER be aware that almost anything connected to the computer is also goinf to carry a ground (printer, monitor, etc) so you'd have to float everything connected to the computer as well. You DO NOT want to float the surge protector that (hopefully) the computer is plugged into or you lose your surge protection.
The problem with floating the ground on the computer is that it's not a safe thing to do. You can try it, though, to see if it helps which at least would confirm that you indeed have a ground loop.
A better solution is to use an isolation device between the scale and the computer.
Ground loops are caused when there's a difference in voltage potential between two locations. Electrical code allows electricians to uses the neutral instead of a ground in electrical boxes, assuming that the neutral is grounded at the breaker box. This can aggravate ground loops.
First off, you may NOT have a ground loop problem. You may have a data drop-out problem. Maybe the signal isn't reliably reaching the computer because it's too weak, or because of poor shielding, or some electrical motor or transformer that's not properly grounded is causing noise in the circuit.
Is this RS-232 serial data? How far are we talking?
Are you using a signal amplfier? RS-232 serial data is limited to a hundred feet without shielding if I remember.
What we used to do was use a device on each end that converted the RS-232 to another serial standard (RS-432? not sure) which CAN be sent very long distances on an unshielded twisted pair with no common ground. Regular phone wire works fine. We've run it as far as 5000 feet.
How far is the computer from the scale? Is there any way, even if only temporarily, to run them both off the same power circuit with a long extension? If it's possible to do this then you can confirm that the problem is indeed a ground loop. If the problem goes away when the scale and computer are on common power then it's almost certainly a ground loop, if not it's something else.
If the above test isn't possible then you'll have to experiment.
Yes you can "float" the ground on the computer and it might help (either cut off the ground prong or use a ground lifter). HOWEVER be aware that almost anything connected to the computer is also goinf to carry a ground (printer, monitor, etc) so you'd have to float everything connected to the computer as well. You DO NOT want to float the surge protector that (hopefully) the computer is plugged into or you lose your surge protection.
The problem with floating the ground on the computer is that it's not a safe thing to do. You can try it, though, to see if it helps which at least would confirm that you indeed have a ground loop.
A better solution is to use an isolation device between the scale and the computer.
---
“The Government of Spain will not applaud those who set the world on fire just because they show up with a bucket.” - Prime Minister of Spain, Pedro Sánchez

“The Government of Spain will not applaud those who set the world on fire just because they show up with a bucket.” - Prime Minister of Spain, Pedro Sánchez

- eGoCeNTRoNiX
- Posts: 7362
- Joined: Wed Oct 23, 2002 12:51 pm
- Location: HELL
Lots of info.
Hey FP,
I've tried all that you've mentioned so far. The computer and scale are less than 10 feet apart. Not sure what the signal is that's being sent, but it's 9-pin if that matters? They are both on the same circuit which is why we thing there's a ground loop. I am gonna get some extra power cables for the printer/CPU/Monitor and try floating them like you said. There's an APC Batter back up on the computers as well.. These are old Celery300s so if something goes wrong it's not a big loss. Though, like others, I'd rather noting go wrong. How would I isolate them? The scale has to input to the computer one way or another, so wouldn't it being connected always cause a ground? Thanx for the info!
I've tried all that you've mentioned so far. The computer and scale are less than 10 feet apart. Not sure what the signal is that's being sent, but it's 9-pin if that matters? They are both on the same circuit which is why we thing there's a ground loop. I am gonna get some extra power cables for the printer/CPU/Monitor and try floating them like you said. There's an APC Batter back up on the computers as well.. These are old Celery300s so if something goes wrong it's not a big loss. Though, like others, I'd rather noting go wrong. How would I isolate them? The scale has to input to the computer one way or another, so wouldn't it being connected always cause a ground? Thanx for the info!
PM before Email People!!
Heat Under eGoCeNTRoNiX
Who Farted? BEANIE!!!
!Welcome to the United States of the Offended!
Heat Under eGoCeNTRoNiX
Who Farted? BEANIE!!!
!Welcome to the United States of the Offended!
- FlyingPenguin
- Flightless Bird
- Posts: 33162
- Joined: Wed Nov 22, 2000 11:13 am
- Location: Central Florida
- Contact:
If the computer and the scale are both plugged into the SAME circuit breaker (so they're on the same AC "leg") and they share a common ground through the outlet, then there can't possibly be a ground loop.
Ground loops only happen when you have two devices with different ground potentials, which usually only happens when you're plugged into different power circuits. Only way you could have a ground loop is if the computer or scale has a defective power supply.
Is it a Toledo scale? It's almost certainly using a 9 pin RS-232 cable.
If you can open the 9 pin connector on the data cable from the scale (it might be a metal connector with screws you can open - otherwise it's molded and you can't get in it) make sure there aren't any broken wires, or any wires touching. Could be you have a short in there.
If I'm reading your post properly you've already tried the next test I can recommend: you've run the bales through with the scale unplugged from the computer and the scale works fine. It's ONLY when the scale is plugged into the computer that it acts up, right?
10 feet away and on the same power circuit, I can't possibly see how it could be a ground loop problem. I also don't see how a signal isolation unit that isolates the ground would help.
For those interested in how to isolate RS-232 optically between two distant devices, this is what we used where I used to work (used one on each end): http://www.telebytedatacom.com/catalog/products/265.htm
For just optically isolating two RS-232 devices, and not worrying about a long cable run, you could use one of these: http://www.peradata.com/products/isolators/ic-232-8.htm
It really can't be a normal ground loop if you're on common power, unless there's a defective power supply or an improperly wired power outlet involved. While using an Rs-232 isolator MIGHT fix your problem (not sure you want to cough up a couple hundred dollars for one though) it won't address the actual problem.
SOME SUGGESTIONS:
- I'm reaching here, but MAYBE the scale is building up a static charge after a certain number of bales go through and when it discharges it scrambles it's little brains. If that's the case then the problem is that the scale is NOT properly grounded. Run a wire from a case screw on the scale to a cold water pipe or something you KNOW is a true earth ground.
- The computer may have a defective PSU and there's AC leaking to the ground.
- Any chance you have another computer you can try?
- Could be a defective power supply in the monitor. I've seen monitors leak RFI (radio frequency noise) so bad that they caused data loss on network cables nearby, and to data cables connected to the computer. The high voltage PSU in a monitor can cause a lot of problems if it misbehaves, and I gather this is probably an old beat-up monitor. Try unplugging the monitor from it's power cord and see if the problem goes away.
- You might also try another serial port on the computer (if there is one). It's a real stretch, but maybe the serial port is screwed up.
- Anyone have a VOM multimeter at you shop? Put it on AC mode at the 100 volt range and measure the voltage between a ground point on the computer and one on the scale WITH THE DATA CABLE UNPLUGGED (touch the probes to a case screw on both units). If it doesn't read anything at 100 volt scale (assuming it's an analog meter an not a digital) try a lower AC scale. If you have a ground loop you'll see some substantial voltage - something over 5 volts. Anything in the millivolt range is meaningless.
- You should check to make sure that the scale and computer (and anything else connected to them) are plugged into PROPERLY wired power sockets. If one socket has the neutral and hot reversed (I've seen a lot of bozos do that in a shop thinking they knew how to wire an outlet and actually being clueless), and your building uses a grounded neutral instead of a true ground, then you'll have major problems. Get one of those $3 outlet testers with 3 lights on it that shows whether the outlet is wired properly and properly grounded. You can get them at Radio Shack or any hardware store. Check all the outlets that the computer, monitor, printer and scale are plugged into and make sure they are the correct phase and grounded.
- Another test you could try: I assume the computer is an ATX case? What happens if you turn off the computer (soft power off, but AC power still connected) and leave the scale hooked up to it? Does it still act up? Some of those scales actually receive data as well as send data (some scales can be programmed from the computer either to enter weight offsets or to upload firmware). Could be the computer is sending gibberish to the scale and locking it up.
- Could be the scale needs a firmware update. Maybe it took a surge and just crashes when it receives a certain amount of data. Some of those scales have a reset switch somewhere.
Hope this helps...
Ground loops only happen when you have two devices with different ground potentials, which usually only happens when you're plugged into different power circuits. Only way you could have a ground loop is if the computer or scale has a defective power supply.
Is it a Toledo scale? It's almost certainly using a 9 pin RS-232 cable.
If you can open the 9 pin connector on the data cable from the scale (it might be a metal connector with screws you can open - otherwise it's molded and you can't get in it) make sure there aren't any broken wires, or any wires touching. Could be you have a short in there.
If I'm reading your post properly you've already tried the next test I can recommend: you've run the bales through with the scale unplugged from the computer and the scale works fine. It's ONLY when the scale is plugged into the computer that it acts up, right?
10 feet away and on the same power circuit, I can't possibly see how it could be a ground loop problem. I also don't see how a signal isolation unit that isolates the ground would help.
For those interested in how to isolate RS-232 optically between two distant devices, this is what we used where I used to work (used one on each end): http://www.telebytedatacom.com/catalog/products/265.htm
For just optically isolating two RS-232 devices, and not worrying about a long cable run, you could use one of these: http://www.peradata.com/products/isolators/ic-232-8.htm
It really can't be a normal ground loop if you're on common power, unless there's a defective power supply or an improperly wired power outlet involved. While using an Rs-232 isolator MIGHT fix your problem (not sure you want to cough up a couple hundred dollars for one though) it won't address the actual problem.
SOME SUGGESTIONS:
- I'm reaching here, but MAYBE the scale is building up a static charge after a certain number of bales go through and when it discharges it scrambles it's little brains. If that's the case then the problem is that the scale is NOT properly grounded. Run a wire from a case screw on the scale to a cold water pipe or something you KNOW is a true earth ground.
- The computer may have a defective PSU and there's AC leaking to the ground.
- Any chance you have another computer you can try?
- Could be a defective power supply in the monitor. I've seen monitors leak RFI (radio frequency noise) so bad that they caused data loss on network cables nearby, and to data cables connected to the computer. The high voltage PSU in a monitor can cause a lot of problems if it misbehaves, and I gather this is probably an old beat-up monitor. Try unplugging the monitor from it's power cord and see if the problem goes away.
- You might also try another serial port on the computer (if there is one). It's a real stretch, but maybe the serial port is screwed up.
- Anyone have a VOM multimeter at you shop? Put it on AC mode at the 100 volt range and measure the voltage between a ground point on the computer and one on the scale WITH THE DATA CABLE UNPLUGGED (touch the probes to a case screw on both units). If it doesn't read anything at 100 volt scale (assuming it's an analog meter an not a digital) try a lower AC scale. If you have a ground loop you'll see some substantial voltage - something over 5 volts. Anything in the millivolt range is meaningless.
- You should check to make sure that the scale and computer (and anything else connected to them) are plugged into PROPERLY wired power sockets. If one socket has the neutral and hot reversed (I've seen a lot of bozos do that in a shop thinking they knew how to wire an outlet and actually being clueless), and your building uses a grounded neutral instead of a true ground, then you'll have major problems. Get one of those $3 outlet testers with 3 lights on it that shows whether the outlet is wired properly and properly grounded. You can get them at Radio Shack or any hardware store. Check all the outlets that the computer, monitor, printer and scale are plugged into and make sure they are the correct phase and grounded.
- Another test you could try: I assume the computer is an ATX case? What happens if you turn off the computer (soft power off, but AC power still connected) and leave the scale hooked up to it? Does it still act up? Some of those scales actually receive data as well as send data (some scales can be programmed from the computer either to enter weight offsets or to upload firmware). Could be the computer is sending gibberish to the scale and locking it up.
- Could be the scale needs a firmware update. Maybe it took a surge and just crashes when it receives a certain amount of data. Some of those scales have a reset switch somewhere.
Hope this helps...
---
“The Government of Spain will not applaud those who set the world on fire just because they show up with a bucket.” - Prime Minister of Spain, Pedro Sánchez

“The Government of Spain will not applaud those who set the world on fire just because they show up with a bucket.” - Prime Minister of Spain, Pedro Sánchez
