Methods of inducing short circuits

[Gabriel Sá Pinto]

Hello everyone, I'm here with a dilemma regarding a company project.
We have a developed software defined radio with 2 separate boards ( one for data, one for power) shielded by a metal case (top, bottom and between boards).

We're on a milestone that requires us to develop test procedures for the delivery of engineering models, and there are several requirements that ask for a short circuit to be triggered purposely either for taking measurements or to test the device's robustness to latching conditions.

However I'm having some trouble finding an efficient way to trigger these short circuits, ideally the unit should be closed but I'm not seeing any way of that happening since specific parts of the PCB must be accessed. Assuming the unit can be opened, what would be the better strategy to trigger a SC?

I've searched for methods and I found break out boxes, however they are most of times wired for specific applications and I'm having trouble finding a configuration that suits our product. Then we have a manual mode of inducing the SC by directly connecting a copper wire between two points or using a set of tweezers, but this can be harmful to the unit or even inaccurate.
Has someone dealt with this kind of situation that can be provide an insight?

Thanks in advance
Best regards,
Gabriel

 

[gophert]

Hello everyone, I'm here with a dilemma regarding a company project.
We have a developed software defined radio with 2 separate boards ( one for data, one for power) shielded by a metal case (top, bottom and between boards).

We're on a milestone that requires us to develop test procedures for the delivery of engineering models, and there are several requirements that ask for a short circuit to be triggered purposely either for taking measurements or to test the device's robustness to latching conditions.

However I'm having some trouble finding an efficient way to trigger these short circuits, ideally the unit should be closed but I'm not seeing any way of that happening since specific parts of the PCB must be accessed. Assuming the unit can be opened, what would be the better strategy to trigger a SC?

I've searched for methods and I found break out boxes, however they are most of times wired for specific applications and I'm having trouble finding a configuration that suits our product. Then we have a manual mode of inducing the SC by directly connecting a copper wire between two points or using a set of tweezers, but this can be harmful to the unit or even inaccurate.
Has someone dealt with this kind of situation that can be provide an insight?

Thanks in advance
Best regards,
Gabriel

When you look at some manufactured PCBs, it looks like parts are missing. These parts were used for testing and development but labeled Do Not Place (DNP) for manufacturing. That said, you could add space for a big MOSFET with low R(on) resistance across your rails to be short circuited. Some rather small TO220 MOSFETS can handle incredible amperage.

 

[mbarazeen]

For the procedure it may need further information such as,

- SC current and power delivered before isolation/ cutoff - design value
- the point where the short circuit is applied, the resistance of PCB, wire etc may play a big role.
- test piece is to be reused or not, since it could be a destructive while testing
- existing protection, line fuse details

a simple method is to make a test circuit to trigger it remotely. the points can be shorted via a high current relay.

 

[Gabriel Sá Pinto]

When you look at some manufactured PCBs, it looks like parts are missing. These parts were used for testing and development but labeled Do Not Place (DNP) for manufacturing. That said, you could add space for a big MOSFET with low R(on) resistance across your rails to be short circuited. Some rather small TO220 MOSFETS can handle incredible amperage.

First of all thank you for your answer.
Unfortunately we can't afford to change the design of our boards, the space is totally maxed out, and no one considered these requirements upon the design stage. Changing now would have consequences far greater than the possibility of not being able to perform this test.
So it is imperative that the method of inducing the short circuit is external to the board rather than "internal".
Regards

 

[Gabriel Sá Pinto]

For the procedure it may need further information such as,

- SC current and power delivered before isolation/ cutoff - design value
- the point where the short circuit is applied, the resistance of PCB, wire etc may play a big role.
- test piece is to be reused or not, since it could be a destructive while testing
- existing protection, line fuse details

a simple method is to make a test circuit to trigger it remotely. the points can be shorted via a high current relay.

Thank you for your answer.

- The requirements where measurements have to be taken are performed with the unit turned off. The robustness tests are done with a nominal operation power of 8.5W (max power during packet transmission) or 6.5W average otherwise.
- The point of application can be in the power transformer as to measure isolation or the signal interface (UART to receiver or structure ground) to test for robustness. There's still another requirement that may require the trigger of the external LCL which we do not control, we're still waiting on clarification for that one. I don't have the resistance of the PCB available at the moment.
- Test piece SHOULD be reused, the purpose is to verify it can survive a short circuit so it must not be destructive. If it happens to be destructed than the test was not passed.
- There are several protections installed OVP and LCL for the various power lines feeding different parts of the board, there's two for separate 3V3 rails, one for 1V5, one for 1V25, one for 5V. One of those protections as mentioned before is before the main primary power conversion stage and it's an LCL external to our company which we don't control.

Regards

 

[ChrisP58]

What is the nature of what the short circuit test is supposed to emulate?

With respect to power. The numbers that you quote are probably your systems normal operating power. What is the maximum possible power/current that the boards and/or the shorting element will see under the fault condition?

 

[ronsimpson]

You are testing the power supply.
I usually find one of the capacitors on the 3V3 supply (ceramic) and add a solder blob across it. Or solder a short wire from one side to the other.

 

[Gabriel Sá Pinto]

What is the nature of what the short circuit test is supposed to emulate?

With respect to power. The numbers that you quote are probably your systems normal operating power. What is the maximum possible power/current that the boards and/or the shorting element will see under the fault condition?

I didn't understand your first question. In one case the unit won't be operating, it's meant for measurement taking. In the other the unit is operating and the UART and PPS (pulse per second) clock interface will be tested.

As for additional specs see my answer to mbarazeen above.

 

[Gabriel Sá Pinto]

You are testing the power supply.
I usually find one of the capacitors on the 3V3 supply (ceramic) and add a solder blob across it. Or solder a short wire from one side to the other.

I believe you're talking a continuous short circuit. I don't know if that'll do the effect I'm after in the case of short circuit robustness. In my understanding the unit would have to be working, than suffer a SC in the referred interfaces, a latch up event would be detected and a recovery would happen, then a subsequent performance test would be performed to assess if the unit remained unarmed.

 

[Cicero]

Sounds to me like you need(ed) to add in some Test Pads. Then you could probe those pads with a bed of nails (pneumatic or manually applied), and simulate/test what you want from those test points.

For (mass) production and testing I've always placed 2x2mm test pads for this reason, with consultation with the test engineers and the pre-written test plan to satisfy all criteria.

For design samples for verification purposes, I dont see a problem with soldering on wires to the necessary points to a manual switch or something, and shorting areas you need to. Documenting all results with scope traces etc. I've done this type of thing for power supply tests, min/max currents, voltage regulation under transition load conditions from min to max, shorts, open circuits etc. This isn't usually necessary for production testing of boards.

 

[schmitt trigger]

Ditto for the test pad approach.

 

[crutschow]

As an alternate to test pads, you could add short test pins (such as one of these) at the desired locations, which will take very little added PCB room.
Those can then just be shorted with a clip-lead jumper as needed.

 

[mbarazeen]

Thank you for your answer.

- The requirements where measurements have to be taken are performed with the unit turned off. The robustness tests are done with a nominal operation power of 8.5W (max power during packet transmission) or 6.5W average otherwise.
- The point of application can be in the power transformer as to measure isolation or the signal interface (UART to receiver or structure ground) to test for robustness. There's still another requirement that may require the trigger of the external LCL which we do not control, we're still waiting on clarification for that one. I don't have the resistance of the PCB available at the moment.
- Test piece SHOULD be reused, the purpose is to verify it can survive a short circuit so it must not be destructive. If it happens to be destructed than the test was not passed.
- There are several protections installed OVP and LCL for the various power lines feeding different parts of the board, there's two for separate 3V3 rails, one for 1V5, one for 1V25, one for 5V. One of those protections as mentioned before is before the main primary power conversion stage and it's an LCL external to our company which we don't control.

Regards

good explanation of overall design.

- since the external main power supply (transformer) is having LCL, you may use its feature to make a disturbance, or momentary shutdown of the power by shorting main input DC bus with a known resistance. you may locate a test point pads or pins for the bus. You have to make sure these test points are under protection zone, By connecting a variable resistor with good power rating, the LCL can be triggered. from the 1st test you can determine a fixed value resistor for test procedure.
- To ensure this test works at different load condition of the unit, maximum value of the resistor can be determined by testing the power supply not connected to the unit/ internal power supplies.
- an intermediate temporary connector (may need a custom made one to be supplied with procedure) can be used to tap the DC input bus from power supply (external supply as you mentioned) if test pads are not found or can not be accessed.
- if the LCL is a self resting one, then duration of latch / interruption has to be determined based on several trials, value of capacitors used at 2nd stage power supplies may influence on the overall behavior of the system. if the duration is fraction of a second then the application may require to use a timer and electronics. (an electro mechanical relay may be slow in response)