Dahua PTZ over-voltage - recoverable?

[Sean Phelan]

I accidentally connected an SD22204T-GN to a 24V (actually, ~26V) supply instead of the 12V it requires. It was only connected for about 20 seconds. Now it seems to be dead - no signals at all on the Ethernet cable.

I can't see anything in the manual about a fuse or circuit breaker. Should I open it up? Is there any change of bringing it back to life?

Thanks all
Sean

 

[jmcu]

Most likely, RIP

 

[tangent]

It's hard to know what got nuked. It could just be some of the components of the power supply / voltage regulators that are internal to the camera or it could have sent way too much voltage to things all over the board. It's partly a question of how quickly certain things got nuked.

Attempting to repair it would be a fairly technical task. If you're hiding an engineering degree or really really want to fix it and have a ton of free time it might be possible. But there would be no guarantee of success. You could easily spend more on tools to try to fix it than it's worth.

 

[Sean Phelan]

Aha. Thanks Tangent; points taken.
I was hoping there might be an internal fuse or well-known component that could be replaced, but apparently not. I do have an electronic engineering degree, and have far too many diagnostic and chip-wrangling tools to hand.... but as you say, attempting to trace the power circuitry would be a massive waste of time with little chance of success.
Thanks again
Sean

 

[tangent]

Aha. Thanks Tangent; points taken.
I was hoping there might be an internal fuse or well-known component that could be replaced, but apparently not. I do have an electronic engineering degree, and have far too many diagnostic and chip-wrangling tools to hand.... but as you say, attempting to trace the power circuitry would be a massive waste of time with little chance of success.
Thanks again
Sean

If you had a functional cam of the same model, it would be a bit easier as you could compare component values, voltages at various points, and waveforms on an oscilloscope more easily. There are most likely test points on the PCB for each of the voltages the camera uses (labeled things like 3V3, 1V2...), I'd at least find and probe those with 12V or PoE applied. It could be as simple as a few blown diodes and maybe a few ldo regulators or much more complex.

The thing that would be the most aggravating would be attempting to source replacement parts for any more complicated components. It can be hard to find let a lone understand the Chinese datasheets.

 

[WA3PNT]

I think that if there is any hope of reviving the Cam, it is if a component very near the input voltage has been compromised. I'm speaking of something like a reverse voltage protection diode, a zener diode, or a current limiting resistor.

These components "should" be easily found by tracing from the point where the 12V input wires connect to the Power Regulating and distribution circuits.

If no damage or failed components are found in this area, then it is most likely that there is damage throughout the circuitry.

Of course there is also the problem of identifying a failed/burned component. With the mass production SMA technology, quite often there are no part numbers on the components themselves. And this is compounded by the variations (US Models, EU Models, China Models, etc.) which quite possibly have different components.

As a Marine with 24 years in the electronic field (now retired since 1984) I've had my share of reverse engineering to keep the equipment on the air.

Good luck in your venture, and please take pictures and keep us posted.

RodeoGeorge

 

[Sean Phelan]

Thanks Guys - this discussion has given me one idea: I'm going to try POE power instead of the 12V feed. As POE is 50V+ it must go through completely different input circuitry. I'll post an update when I've tried that, with good or bad news.

 

[awsum140]

Don't wan tot rain on your parade, but the PoE power is converted to 12vdc to run the camera.

 

[Fastb]

When I try to diagnose a failed pcb, I look for a fuse. Then for signs like burnt trace, burn component, etc. Lastly, the "sniff test". This often identifies what part got fried. Only then do I pick up tools....

 

[tangent]

When I try to diagnose a failed pcb, I look for a fuse. Then for signs like burnt trace, burn component, etc. Lastly, the "sniff test". This often identifies what part got fried. Only then do I pick up tools....

100%. The first tools I'd reach for would be light and magnifiers then a stereo microscope if available or even a USB microscope.

 

[tangent]

Don't wan tot rain on your parade, but the PoE power is converted to 12vdc to run the camera.

yep, and they aren't as isolated as you'd hope. when i took a meter to the barrel jack of a dahua camera with PoE applied there was 1.2v on the jack.

 

[awsum140]

I lost a Granstream camera when I accidentally used a crossover cable on PoE. No isolation or protection that worked in that case.

 

[Fastb]

when i took a meter to the barrel jack of a dahua camera with PoE applied there was 1.2v on the jack.

Interesting.
Where did you put the other meter lead? 1.2V from the barrel of the empty barrel power connector to what? Camera chassis ground?
The high input impedance of the voltmeter means you can get erroneous readings. If you put a load on that 1.2V, such as a LED and current limiting resistor, the voltage might disappear.
So an isolated DC input gnd (ie: the barrel) on a poe powered cam may have acceptable isolation, right? You might see the same 1.2V on the center contact (the V+ side of the barrel connector). Again, in reference to what true ground you are measuring from.
Seeing voltage with a high input impedance meter can have other explanations.
In house wiring, when I put a meter on black or white conductors, to check for voltage, I will get readings when the circuit breaker is off. Not 120V AC. Which confirms the breaker is off. (just doing a safety check). But sometimes tens of volts. I chalk this up to coupling to live cables (on other circuits), or the long runs on the "turned off" circuit acting as an antenna picking up RF. When I bridge to white & black with my fingers, the voltage reading drops to zero.

@tangent : Thoughts? Maybe clarification on the setup of your measurement?

 

[tangent]

@tangent : Thoughts? Maybe clarification on the setup of your measurement?

I measured both parts of the barrel jack, the center pin and the outer ring. no load. just a cheap multimeter.
I debated the accuracy of the measurement too, but it goes away shortly after unplugging the camera once the caps discharge.

In house wiring, when I put a meter on black or white conductors, to check for voltage, I will get readings when the circuit breaker is off. Not 120V AC. Which confirms the breaker is off. (just doing a safety check). But sometimes tens of volts. I chalk this up to coupling to live cables (on other circuits), or the long runs on the "turned off" circuit acting as an antenna picking up RF. When I bridge to white & black with my fingers, the voltage reading drops to zero.

I don't think I've seen tens of volts doing that, typically <5v. How close is the nearest radio tower

 

[TonyR]

.......The high input impedance of the voltmeter means you can get erroneous readings. If you put a load on that 1.2V, such as a LED and current limiting resistor, the voltage might disappear. ........ In house wiring, when I put a meter on black or white conductors, to check for voltage, I will get readings when the circuit breaker is off. Not 120V AC. Which confirms the breaker is off. (just doing a safety check).

I read this and my hair stood up. I was a traffic signal technician for over 31 years. I'm sure by my avatar you'd never guess! LOL

Anyway, I used either a Triplett 630 or a Simpson 260 (my fav), both analog, low impedance VOM's to troubleshoot 120 and 240 VAC issues in the field. Such a VOM will show things, USEFUL things, a digital meter cannot ( at least, back then) and the user needs to know what to look for.

When the traffic signals are in flashing mode, with an analog VOM you can read an induced voltage from 5 to 10 volts on the non-flashing colors (green or yellow when it's an all-red flashing intersection), up and down, from 0 to 10 VAC, at the same rate as the flasher (40 to 50 times per minute) but NOT on any wire that is shorted to ground, say by an underground failure, vandals (morons would unscrew lamps and put penny in socket then screw lamp back in) or other issue.

On an open or 'cut' wire, a wire supposed to be connected to between 2 and 4 total 150 watt incandescent traffic signal lamps or 25 watt LED modules, you would get a voltage swing as much as 100 VAC because there was no load, no resistance, because the wire was 'cut' (open) before it made to the lamps, through the lamps and then to neutral.

A digital meter would never see either of those indicators...most of the time, the sampling rate wherein it would sample the analog then convert to digital to produce the reading was too low in early digital meters. By the time it 'saw' the voltage, decided what to display and displayed it, it was gone. And the digital meters had too high of an input impedance to be used on trying to determine what's happening in the field when you have 5 or 6 runs of PVC conduit underground, each with 12 to 48 each of #14 copper conductors running in parallel, right next to each other, actually intertwined to a degree. Being able to read how a conductor reacts to changing induced voltages will tell you if the conductor is OK, open or shorted....if you know how to read it. Maybe today's digital meters can tell you ( I retired in '04) but the ones around from about '80 to '04 sure could not.

I had a digital meter and used it a lot when doing on/off, go/no-go testing or locating wilres via continuity but the meter I used to tell me what was REALLY going on in the field (open wires, shorted wires, etc.) was my trusty old analog VOM.

Yep, I have become my dad.

 

[tangent]

@TonyR I can almost picture your dad telling the story, "when I was a kid we had to walk 6 miles, uphill both ways, through the snow to get to school and we couldn't just buy an electric meter. We didn't even have electricity. The closest I got to a 'wheatstone bridge' was crushing wheat between 2 stones while standing on a bridge."

 

[TonyR]

@TonyR I can almost picture your dad telling the story, "when I was a kid we had to walk 6 miles, uphill both ways, through the snow to get to school and we couldn't just buy an electric meter. We didn't even have electricity. The closest I got to a 'wheatstone bridge' was crushing wheat between 2 stones while standing on a bridge."

LOL! Don't forget "..... had newspaper in my shoes because of the holes".

Ah, yes...visions of SNL's Dana Carvey as "the grumpy old man": "Toys? We didn't have toys, we played with rocks.. AND WE LIKED IT!"

 

[geezer]

"Being able to read how a conductor reacts to changing induced voltages will tell you if the conductor is OK, open or shorted....if you know how to read it."

Sometimes I can't remember to put on my pants but I have good recall of my time as a repairman for Ma Bell in the 70's. We were provided with sturdy and very sensitive analog VOM's.

Looking for an open could sometimes mean going from pole to pole, in the direction of the central office, looking for dialtone. Not fun at all, especially in the rain, especially if it meant climbing bare poles with cleats.

The VOM could read capacitance and had a button that would send a small pulse. With enough experience at watching the reading peak and decay, you could make a decent guess at the distance to the open.

 

[TonyR]

The VOM could read capacitance and had a button that would send a small pulse. With enough experience at watching the reading peak and decay, you could make a decent guess at the distance to the open.

Exactly!

Neither the Triplett 630 or the Simpson 260 had a setting to expressly check capacitors but I'd put the dial on R x 1K (or 10k depending on the cap rating in uF) and check large filter capacitors in linear power supplies or smaller caps in Schmidt trigger timing circuits; you'd charge it up with the + and - leads of the meter on the cap one way, then reverse the leads and watch the charged cap decay. It wasn't an accurate test but experience would allow you to make a pretty good call that a 1,000 uF cap or a 100 uF was discharging too quickly or not storing a charge at all as it should. Only a low impedance meter would allow you charge up a cap with the meter's batteries like that (batteries used in the meter's internal Wheatstone bridge to calculate the resistance under test).

I love digital radio tuners, digital watches ( analog's OK, too), all kind of digital displays....but given a choice of only one meter I'd take a good analog VOM over digital any day.

 

[WA3PNT]

I think we have strayed a bit from the subject of this thread.

RodeoGeorge

 

[TonyR]

I think we have strayed a bit from the subject of this thread.

RodeoGeorge

Not totally. The OP brought up a subject which required troubleshooting voltage issues with a camera. A discussion insued regarding how to interpret readings and how to discover problems using a meter, VOM, etc.

That being said, feel free to change the topic, contribute...or go to another topic.