repairing a remote voltage monitor circuit.

if there's only a few cheap chips in there replace them.

if it was a lightning strike the bits that would have got hit the hardest are those nearest the cables, at both ends of the cable. so check the circuitry behind the meters and that the meters work...

damaged parts aren't always visibly damaged.

Bye. Jasen

Here's what I'd try. First, the meters can be loop powered (two wires) or locally powered (four wires) - look at the meter connections. If they're supposed to be locally powered, make sure that power supply is working. It also wouldn't surprise me if this setup brings +24 V from the top of the hill to power the meters separately from the signaling loop; if the +24 V doesn't make it down there, the meters won't work.

If that checks out, then: Assuming it's 4 to 20 mA signaling, get a 24 V power supply or 24 V worth of (small) batteries. Connect a 1 K resistor and 5 K pot in series, and connect this across the 24 V supply. You can now draw 4 to 24 mA (at 24.0 V nominal) from the supply. (If you don't like going over 20 mA, use a 1.2 K resistor.) Go to where the meters are, disconnect one of the meters from the signal wires that go up the hill, and put the meter in series with the pot, resistor, and batteries. The meter will have some internal resistance, so you won't be able to go to the full 24 mA, but you should be able to vary the reading on the meter by adjusting the pot. If you can't make the meter respond, then there is a fault in the meter.

If you can make all of the meters move with your test rig connected directly to the meters, try the same test, but from the top of the hill. Disconnect the signal cable from the electronics up there and use your battery-resistor-pot rig to send a varying current down the cable. If the meters don't move, suspect the signal cable. If they do move, you probably won't get exactly the same readings as you did before, due to the added resistance of the signal cable, but you should get some kind of varying reading.

If all of that checks out, you're left with the electronics at the top of the hill. You might check the input wires from the battery bank and the shunt to make sure the right signals are getting to the box. If that's OK, then connect your 1 K resistor across the electronics box in place of the wires to one of the meters, and measure the voltage across it to see if the box is trying to drive the meter. If you can, put a load on the batteries to see if the voltage across the 1 K resistor changes. If you can't get a reasonable output from the electronics, I would then vote for taking the electronics box home where you can test it at your leisure.

I wouldn't worry about getting shocked as much as I would about shorting the output of the battery bank. It's possible to shock or burn yourself on 24 V but you have to work at it a bit. It's pretty easy to melt a wrench by dropping it across the terminals of a battery. If there are no fuses in the system, now would be an excellent time to add some.

Matt Roberds

They are loop powered, according to the manual I have.

Thanks, I'll try this.

I've tested all these cables for continuity, and they're OK.

Thanks for the detailed reply.

andy.

--
http://www.niftybits.ukfsn.org/

remove 'n-u-l-l' to email me. html mail or attachments will go in the spam
bin unless notified with [html] or [attachment] in the subject line.

Why? It was working perfectly well before it broke, and the signal wire is still in place. (If it's damaged we can use a different set of wires, or install a new one which they have anyway). so as far as I can see there's no need to rebuild the whole thing, which is adding extra complications and costing them a lot in parts at the same time.

--
http://www.niftybits.ukfsn.org/

remove 'n-u-l-l' to email me. html mail or attachments will go in the spam
bin unless notified with [html] or [attachment] in the subject line.

put a wireless system in.

--
Real Programmers Do things like this.
http://webpages.charter.net/jamie_5

On Sat, 22 Oct 2005 04:23:25 GMT, snipped-for-privacy@worldnet.att.net put finger to keyboard and composed:

I'd leave the rig at the bottom of the hill. At the top of the hill I'd put a 100 ohm resistor between the two ends of the signal cable. That would enable me to detect shorts in the cable, and to watch the meter while I'm adjusting the loop current.

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

On Fri, 21 Oct 2005 19:52:01 +0100, Andy Baxter put finger to keyboard and composed:

It may help us visualise your setup if you described the various chips at both ends, particularly the interface ICs. One current loop implementation that gave problems for me involved a 7406 open collector driver at one end and an optocoupler at the other.

I'd be interested in knowing what circuitry is between the meter movement and the interface. Does the current loop interface chip merely translate 4mA and 20mA currents to 0 and 1 logic levels? Does the meter operate in PWM mode, ie does it average the current pulses to get an analogue reading? Does it possess any "smarts"?

What do you see on the meter if you drive it directly from the diode range of a DMM? My DMMs produce 1mA on this range. You could use a second series connected DMM to measure the current. Or you could replace the meter with your DMM and measure the loop current directly.

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

Next time a lightning bolt strikes that hilltop,a wireless system will survive a whole lot better than the wired system. (especially if the aerial is mounted low) fibre-optic is another option.

it's a hill top, they get struck quite regularly.

It all depends on how often you get electric storms in that area, and how often you want to be repairing that device.

Bye. Jasen

There are some pictures at

p1010087.jpg is a photo of the main circuit board from the top. The main part of the circuit is two AD694 current loop drivers, each fed by a IMA118 instrumentation amplifier - one for the voltage measurement and one for the current shunt.

I think I can probably work it out - I just got spooked by the thought of having to reconstruct the circuit diagram from the circuit itself, and I've never done any diagnostic / repair work before, so I'm not sure what's the best way to get started on something like this.

It looks like the bloke who built it should be sending me a circuit diagram, so it shouldn't be as hard as I thought.

It's just a standard 4-20 mA current loop. Not PWM, looking at the datasheet for the driver. The meter circuitry just has a couple of resistors and a diode on each meter. The meters are driven directly from the current loop, not by any active circuitry.

Don't know, but I'll try some test like this next time I go up.

Thanks for your help - I'll post again to say how I get on with it.

--
http://www.niftybits.ukfsn.org/

remove 'n-u-l-l' to email me. html mail or attachments will go in the spam
bin unless notified with [html] or [attachment] in the subject line.

with the work they put into the manual, you would have thought they could had made a real board or at least a hand sketched one. oh well. in any case, i would suggest looking at the op-amps and input devices. and even the regulator.

By looking at ICs only, I believe this is a voltage and a current or a two channel voltage monitor. I suspect that two

4-20 cables carry analog data down the hill.

Two inputs would be the INA118 Op amps. Two 4-20 ma wires would be driven by pin 11 of the AD694.

Step one: what does AD694 pin one connect to? If to pin 2 or 7, then 0 to 10 volts on pin 3 cause the 4-20 ma output. IOW make changes to the circuit board voltage input (to inputs of INA118) and observe corresponding voltage changes on pin 3 of AD694. Now you know the input is good (or bad).

Step two: see what comes out of AD694. Temporarily replace the cable going down hill with a 100 ohm resistor in series with a digital multimeter in 200 ma range. As those inputs to INA118 change, then current measured by multimeter will change from 4 to 20 ma.

Step three: Assuming the above tests have worked, then perform the same step two test, this time by replacing the 100 ohm resistor (even sold in Radio Shack) with the cable going downhill. IOW restore the connections except place the multimeter in 200 ma mode in series with one wire of that downhill cable. Again, changes to the input should measure same 4-20 ma numbers on meter.

Step four: report back everything no matter whether problem is solved or not. This is a two way street. When done, all should have learned something.

Step five: warning - the most common way to cause meter damage is to leave it in current mode when done. As a rule, always switch meter back to voltage measurement after taking current. Just a good habit to get into.

Andy Baxter wrote:

Thanks - I'll let you know how I get on.

--
http://www.niftybits.ukfsn.org/

remove 'n-u-l-l' to email me. html mail or attachments will go in the spam
bin unless notified with [html] or [attachment] in the subject line.

On Sun, 23 Oct 2005 15:28:43 +0100, Andy Baxter put finger to keyboard and composed:

I found this datasheet:

It looks simple enough to me now. Until I saw the photo I was thinking that the flashing LEDs may have been driven by a microcontroller as part of a POST. The 4mA-20mA current loop threw me as well, until I saw that the 4mA offset could be switched off at pin 9 of the AD694. A

0-20mA current loop makes more sense than 4-20mA when driving an analogue meter.

I suspect the LEDS may be connected to the ALARM pins (10) of the AD694s. The alarm goes low when the output voltage (11) has run out of headroom, ie when Vout = Vsupply - 2V. This could occur if the IC is driving a loop with an excessively high resistance. I'm guessing that the 555 timer may be wired as a one-shot, and that it applies a momentary test voltage to each of the AD564 inputs via the relay. Or maybe it just momentarily disconnects the current loops. This could explain the flash of LEDs at power-on.

It seems to me that the easiest test would be to substitute your DMM (in ammeter mode) for the load at various points in the circuit, firstly at the output of the PCB, then at the end of the cable.

Just one other thought. Does the meter display 0-30V, say, or does it use an expanded scale, say 20V-30V? It seems to me that anything below about 20V (for a 24V battery) would be uninteresting. I reckon it would be much better to see the detail at the top end of the voltage range.

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

From the manual, it's a 4-20 mA loop, but I don't know how this works at the meter end.

That makes sense. There's also another timing element though - the system was designed to operate a solenoid to turn the inverter on and off remotely. This worked by breaking the circuit of the voltage measuring current loop, and triggering a deliberate fault condition. I think he did it that way to save on wires. That has some kind of timing associated with it, according to the manual, so the 555 might be for that.

It's using an expanded scale - something like 20-30V but I'm not sure.

--
http://www.niftybits.ukfsn.org/

remove 'n-u-l-l' to email me. html mail or attachments will go in the spam
bin unless notified with [html] or [attachment] in the subject line.