Safety rated voltage monitor circuit 2

[kedminster]

All:

I am designing a test system that involves a DUT (device under test) that contains a capacitor bank (5000uF @ 600 VDC). Our test system contains a bleed resistor (7.5k ohm) to discharge the cap bank at the end of the test. I need a safety rated circuit that will detect the voltage of the capacitor and output a safety rated (force-guided relay, etc.) signal when the cap bank voltage is below some threshold (say 30V DC) to allow the test system enclosure latch to release.

I could certainly design a comparator circuit to do this but would rather buy a safety certified device/module that is specifically designed for this sort of application. Does anyone know of such a device?

We will also require the test operator to use the proper PPE and use a secondary method (DMM, etc.) to verify the discharged bank prior to handling the DUT. Additionally the bank does have integral bleeder resistors to ensure it stays discharged after being removed from the test.

Thanks in advance for your assistance/suggestions!

Karl Edminster

http://www.electromechanica.com

Does that smell like burning FR-4? Nah...just the FETs breaking in...

 

[VEBill]

An alternate interlock approach that is often used is to have a fail-safe switch that will cut power and (in this case) apply a fast-acting bleeder to the DUT.

Using the above approach as the ultimate operator safeguard might allow you to add a mechanical lock (of the sort you described) using systems and components that are not certified.

Disclaimer - I'm not an expert in the regulations that might apply.

 

[itsmoked]

I believe you will need to roll your own. I have never heard of, nor seen such an "approved" device.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[jraef]

Siemens makes a safety rated "Standstill Monitor" that monitors the decaying voltage of a motor after being switched off to determine when it has come to a stop. So it is essentially a "Voltage Monitor" in that sense, but it is not specific to your application and I don't think any safety regulators will accept it being used for anything not part of its primary purpose.

But in essence, you may not really need the SENSOR device to be safety rated, as long as it feeds a "Safety Monitor" relay that then controls the access. So from that standpoint, any relay that holds in at a threshold level and drops out below that would do the job. You can get Voltage Monitor Relays that have over-shoot / under-shoot functionality. This one from Siemens (I used to work for them, it's what comes to mind first) has a version that works up to 650VDC (the 3UG4632-1AA30), but I'm sure there are plenty of others.


Also, check this out as a way to verify without opening the door, they are rated for up to 1000V AC/DC, you would put them on the door and use a non-contact voltage detector to see if there is still voltage on them (although you won't know the level):

I've been using them on Solar combiner boxes recently, it's a great way to test w/o needing PPE.

Alternate version; LEDs that flash when a voltage is present, 30-1000VDC


(click images to link to their websites)




"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[kedminster]

I want to thank everyone who responded.

VE1BLL- Good suggestion but it would be hard to discharge this bank fast enough to make it safe in the time it would take a user to open the enclosure. The other problem with that approach is the stress put on the caps during a high current discharge and the greater potential for residual voltage due to dielectric absorption. Slower discharge tends to reduce this effect

Itsmoked- Yep, I am going to have to build this from scratch... I agree with your assessment

jraef- Great suggestions. I had seen the voltalert LED modules in the back of EC&M but thanks for jogging my memory. I had no idea they would be useful for DC but the datasheet clearly shows they are. Going to order one today...

Question: How do you use the non-contact pre-test points with DC? I don't know of a non-contact sensor that would be able to tell that there is a DC potential behind an insulated barrier unless there was current flow (and resulting magnetic field). If the wire dead-ends at the pre-test point there shouldn't be any current flow so there wouldn't even be a static magnetic field. How do you use these on DC voltages in your PV combiner boxes? What sensor do you use?

Karl Edminster

http://www.electromechanica.com

Does that smell like burning FR-4? Nah...just the FETs breaking in...

 

[jraef]

Huh, hadn't thought about that. I just saw that they rated it for 1000VDC, didn't think it through all the way but you're right, you would need a tester capable of that. There are some non-contact testers that work on DC (don't ask me how though), but they apparently stop at 250VDC, which is all that the one's I used them for needed.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[MacGyverS2000]

jr,

I've never seen those panels with the non-contact points on them (of course, power isn't my field)... very cool.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[VEBill]

5000uF isn't really that high a capacitance (even taking into account the 600 volts).

100 ohms, peak current 6A (but dropping rapidly), only about 1.5 seconds to reach 30 volts. Although the intial peak power might be 3.6kw, the extremely low duty cycle (*) and some reasonable thermal mass would probably allow use of a small power resistor (xx watts).

(* it's a ultimate safety system; the operator shouldn't be using 120 times an hour.)

The voltage will not rebound since the discharge circuit would remain connected for 'a while', plus considering the built-in 7.5k bleeder you mentioned.

It couldn't be simpler. A small power resistor and an interlock switch.



The basic concept could be optimized by using a simple (approximate, one big power transistor) version of a constant current sink instead of a resistor. This would allow you to further limit the peak current at the start, while also hastening the discharge to reach 30 volts. A flat discharge current instead of an RC discharge curve.

Further system architecture optimizations might include one slower (maybe 5 seconds) discharge circuit automatically applied at the end of the test, and another faster (safety) discharge circuit applied when the cover is opened. The first would include an indicator, and the second circuit might include a nasty buzzer to discourage use.

SCRs could be used if you want the cover interlock switch to last "forever".

You should also check the tolerance of the capacitors. Worse case maximum values should be considered. Some capacitors are quite wide tolerance on the plus side. That needs to be taken into account.

 

[kedminster]

I really appreciate all the great feedback!

jreaf- Can you send me the manufacturer/model of the non-contact DC voltage sensors you use? I would like to learn more about them.

VE1BLL- Good point on the cap bank tolerance! These big electrolytics are +/-20% Additionally there is one point in the test where we go to 780VDC for an overvoltage test. The test system power supply is capable of 800V so I really should use that instead of 600V for my calc's

Thanks for your good suggestion for the rapid discharge. I agree with your proposal for a discharge circuit but what I really need is a safety rated sense circuit that senses that the voltage really has been discharged. The methods you described do not provide a safety rated feedback channel. While I would agree that the discharge plans you have described are pretty reliable I need to come up with a secondary method of testing the voltage across the cap that is safety rated and does not require human intervention/interpretation (i.e. meter).

Unfortunately I have to take my engineering hat off and put my attorney hat on and think of how I would provide protection against 'reasonably foreseeable' failures of the safety system and do a FMEA. I know it stinks that we live in such a litigious society that keeping the ambulance chasers at bay trumps people being smart when working with dangerous equipment but it is what it is...

Karl Edminster

http://www.electromechanica.com

Does that smell like burning FR-4? Nah...just the FETs breaking in...

 

[VEBill]

The way we ensure 'ultimate HV safety' is provide a grounding stick.

The operators are trained in the normal UUT post-test shutdown processes, and then (to make sure) they poke the +V with a probe grounded to chassis and earth grounds.

If everything is working properly, then nothing will happen because the PS is already at zero due to cut-power, the bleeders and mandated wait times.

But if there's been a series of failures and mistakes, then they'll get a big frightening spark. So they'd then lock-out the bench, change their underwear, and call for engineering help.