Training board DC circuit protection suggestion

[eagle1109]

Hi,

I've been working on repairing this power electronics training board for several months, I changed series resistors, voltage regulators, H-bridge MOSFETs, the power transistor and Op-amps.

So apparently, the faults are happening within the DC circuits, there are AC circuits, which are the 1-ph and 3-ph but they are protected with a fuse for each one. And the fuses work fine, there haven't been any problems with the AC lines, the fuse get broken every time and we change it and problem solved, but the real problem in the DC circuits.

And I just realized that this afternoon and I thought why can't I include a protection component, and save myself the time of repairing the boards every time ?

This is the board:

So, for 1-ph -> F1 and 3-ph -> F2. But the DC power supply has no protection. Most the problems comes from +10/-10V.

This is the schematic for the power supply part:

So, the problems happen with the +10/-10V and +15/-15V.

And this is what I have to deal with !

The burns eat the PCB, so I might be able to fix it now, but next time the PCB would be harder.

So I thought of finding a way to protect the DC main lines. What you think of ways to do ?

I thought of including a circuit with a RED LED and a buzzer, so when the trainee do a mistake and short something, the LED lights up and the buzzer rings so one can undo this particular experiment wiring.

 

[BradtheRad]

Not sure if this is the problem... you might need to make your N2 regulator of the type intended to be a negative supply rail. Right now it's identical to N1 (LM1085) which is designated a positive regulator IC.

Your N4 (7910 IC) is properly chosen since it's the negative counterpart to your N3 (7810) regulator.

The 7800/7900 series IC's are built with automatic thermal shutdown. This should provide a measure of protection just like a fuse.

Voltage Regulators 78xx and 79xx Family specifications and uses

Voltage Regulators- The Linear voltage regulators are components with or without associated circuitry which provide a stabilized output...

www.electroniclinic.com

So is it that big a difference? Your system works most of the time. However there seems to be unexpected situations that break your system. If positive regulators were sufficient for every situation then why make negative regulators?

 

[eagle1109]

Not sure if this is the problem... you might need to make your N2 regulator of the type intended to be a negative supply rail. Right now it's identical to N1 (LM1085) which is designated a positive regulator IC.

N1 and N2 as I understood with the way that V1 and V2 are done, it's like a center tapped transformer to provide +/- voltages.

So N1 is connected in away to maintain the +15V and N2 with the way of V2 and the caps around it is actually maintaining -15V.

I think this type of regulators work either for positive or negative voltages. I don't know if the LM1085 is intended for only positive voltages or both positive and negative. But the question here is why this company did it this way ? it seems working, it's a training board and I just started working on fixing the problems and get to know the how the board is designed.

Your N4 (7910 IC) is properly chosen since it's the negative counterpart to your N3 (7810) regulator.

The 7800/7900 series IC's are built with automatic thermal shutdown. This should provide a measure of protection just like a fuse.

Voltage Regulators 78xx and 79xx Family specifications and uses

Voltage Regulators- The Linear voltage regulators are components with or without associated circuitry which provide a stabilized output...

www.electroniclinic.com

I got several fried 7910 and 7810 regulators, so my analysis is that they took a big current draw through them to things more than the maximum current, because there's no current protection.

So is it that big a difference? Your system works most of the time. However there seems to be unexpected situations that break your system. If positive regulators were sufficient for every situation then why make negative regulators?

It's working out of the box, but our problem is that during the training, students do the wiring mistakes in many ways. I don't know exactly what the wiring mistakes but for sure they shorted something to the ground or connected a high voltage from +10/-10 rails to a low voltage point.

I can't work on fixing these boards all the time, it's time consuming and the boards won't be fixable in the future.

So I want to do something in these boards to prevent the problems.

I've got an idea, I might contact the company that designed theses boards and search with them about a possible solution.

--- Updated Feb 23, 2023 ---

I just sent them an email with couple photos of fried components.

 

[BradtheRad]

It's obvious to install a fuse on every DC supply, but uneconomical since they'd go rapidly. Circuit breakers are an alternative however low voltage DC models do not appear readily available.

The manufacturer's motivation to sell training units. Be surprised if he feels like installing all kinds of safeguards (and raising each unit's price).

Suppose you install a relay in the power path? Arrange for an excess of current to activate the coil quickly. It should be possible to wire the coil so it stays energized indefinitely after the contacts close.

Or a latching relay? A latching relay activates from a single pulse through the coil. Then it's released by a pulse in the opposite direction.

----------------------------

The boards appear to have jacks for banana plugs (labelled X35 etc.). Are these out in the open? Do trainees hook things up directly to the boards? Or can they only access the DC supply terminals?

The diagrams indicate a multitude of functions in these training units. It's a lot for a trainee to learn immediately. Mistakes are easy. Spectacular too, when there's smoke and burning odor. Such embarassing episodes can discourage a student from learning further.

I thought of including a circuit with a RED LED and a buzzer, so when the trainee do a mistake and short something, the LED lights up and the buzzer rings so one can undo this particular experiment wiring.

This sounds ideal of course. The student needs to understand what causes what. As the relay opens it could send power to your alert circuit. An RS flip-flop is sufficient to keep it activated until you tend to it.

 

[betwixt]

I think I would tackle the problem in a different way. The +/- 10V rails come from a 1A regulator and all the current flowing through it comes from a 5A regulator. I would put a current limiting stage between the rectified AC and the 10V regulators, bypassing the 15V regulator altogether. You could use another 7810 and 7910 in constant current mode ahead and set to whatever maximum current you anticipate being enough. The characteristic of the 10V rails would then to be constant voltage up to maximum current then to limit at the set current and still keeping the thermal shutdown as well.

It would require some minimal adjustment to the rectifier end to get the polarity with respect to ground.

Alternatively, you could try replacing the 10V regulators with a 'wrap around' bypass transistor and bias clamps, it might be easier to implement on the existing board but the penalty is some extra voltage drop in the clamp and the existing drop from 15V to 10V may not give enough overhead for the regulator to work properly. Something like this but adapted for 10V:

DC Voltage Regulator Circuit: How to Increase Current 7805 Voltage Regulator CIrcuit

This page is about an easy way to improve the current 7805 voltage regulator circuit using an external NPN transistor. It is tailored to the needs of the load current.

dc-voltage-regulator.blogspot.com

Brian.

 

[eagle1109]

It's obvious to install a fuse on every DC supply, but uneconomical since they'd go rapidly.

Yep, the fuses in this board are for the 1-ph and 3-ph lines, and we use them a lot during the training sessions. But the good thing is that the stuff along the AC circuits are protected. Also for the 1-ph the manufacturer suggested 1.25A but I'm using 0.5A for more safety. Maybe after protecting the DC circuit, I might go back to use 1A and check if everything is good. Because I'm still learning and exploring this board, but I think what is most important now is to put everything in the protection and safety of the board, then the trainer and trainee can investigate and learn more in the experiments.

But with problems of burning things in the board, it would be difficult for the trainer to go on and control the experiment session when one of the trainees has done a mistake and fired something, now this trainee can't complete the experiment and then this bench isn't fully ready to do this experiment and maybe later experiment. So it's annoying for the trainer, I've worked in this course for couple years, but other trainers now are working in it and I'm charged to do other stuff in the department like maintenance. So this board will cost me a lot of repairing time, so instead of that I thought of doing something to prevent any shorts in the future and then I will have less work pressure and more time to do other responsibilities.

Circuit breakers are an alternative

Suppose you install a relay in the power path? Arrange for an excess of current to activate the coil quickly. It should be possible to wire the coil so it stays energized indefinitely after the contacts close.

Or a latching relay? A latching relay activates from a single pulse through the coil. Then it's released by a pulse in the opposite direction.

I like this idea. Yep, there are 12V and 24V relays in the market.

I'm thinking of a sensing circuit for a specific current value; e.g. 1A, there are many current sensing circuits online, then provide that amount of current with suitable resistors to turn ON/OFF a transistor to energize the trip relay for either the mains voltage 220V or the whole DC block with a RED LED and a buzzer.

I don't know if this value is safe or not because there are different parts for the DC circuits, some might take 1A no problems and some parts might get burned.

I have the full schematic file from the company, I will put it in the attachment.

however low voltage DC models do not appear readily available.

Why ? which ones you mean ?

The manufacturer's motivation to sell training units. Be surprised if he feels like installing all kinds of safeguards (and raising each unit's price).

Didn't get this point clearly, so if the manufacturer put a good protection system and raised the unit's price, where is the problem then ? I think the raise will be little not much. I don't know I have no much experience with the business and marketing stuff.

The boards appear to have jacks for banana plugs (labelled X35 etc.). Are these out in the open? Do trainees hook things up directly to the boards? Or can they only access the DC supply terminals?

Yes, these boards use smaller version of the normal banana cables. This picture is from their website.

https://www.hera.de/en/electriceducation/products.html

The diagrams indicate a multitude of functions in these training units. It's a lot for a trainee to learn immediately. Mistakes are easy. Spectacular too, when there's smoke and burning odor. Such embarrassing episodes can discourage a student from learning further.

Totally agree with you.

This sounds ideal of course. The student needs to understand what causes what. As the relay opens it could send power to your alert circuit. An RS flip-flop is sufficient to keep it activated until you tend to it.

Absolutely on point. I'm starting to design a circuit and do some breadboard experiments.

--- Updated Feb 24, 2023 ---

Here's the schematic file.

The +/- 15V rails supply power the different op-amps, other ICs, other electronic sub-circuits, power transistors and H-bridge MOSFETs.

So, yes it might be more safe to bypass the 15V from the limiting stage, because this voltage supply different circuits.

I still like the idea of a latching or trip circuit with a LED and a buzzer to alert both the trainer and trainee that there something wrong in the experiment wiring.

 

[eagle1109]

I have a quick question. I searched about the bridge rectifier and it's rated for 8A/600V, isn't that a lot for this circuit since the input voltage won't exceed 15V/AC ?

 

[betwixt]

Yes, but as they are mass produced it may be cheaper than a low voltage one. Performance will be the same.

Brian.

 

[eagle1109]

I started to design the circuit on Proteus, but I didn't get -15V !

What I'm missing here ?

 

[betwixt]

To do it that way you need two independent AC sources. On the original design they came from different transformer secondaries.

Brian.

 

[eagle1109]

Same problem:

 

[FvM]

The latest circuit is o.k. and outputs -15V, as intended. Voltage difference between ground and -15V terminal is nevertheless +15V.

To measure negative output, you'll connect voltmeter negative terminal to ground, not to -15V output.

 

[eagle1109]

Yep, I forgot to flip the voltmeters

--- Updated Feb 25, 2023 ---

What you think of this module ?

https://www.aliexpress.com/item/10000185833022.html?gatewayAdapt=glo2ara

--- Updated Feb 25, 2023 ---

Found this circuit using op-amp:

 

[betwixt]

The current transformer unit is a good choice if it offers a manual reset. If you use it to cut the supply when too much current flows, it will immediately sense no current and reconnect again in a continuous loop.

I think ideally you need something that senses over-current and shuts the power off until the problem is solved and it can be manually turned on again.

The op-amp schematic will work but isn't easily adapted to monitor negative and positive supplies and I suspect your overload could occur on either or both. It also suffers the cyclical tripping problem.

Brian.

 

[FvM]

Do you exactly know which experiment setup (mis-wiring, overload or short circuit) caused the shown damage? The IC voltage regulators are internally protected against overload and short circuit, but not against externally applied voltage. Unfortunately there are many ways to do so in a wrongly setup power electronics experiment.

Before designing individual protection means I would sketch the expected overload/ damage cases and consider how they can be prevented, respectively if at all.

 

[eagle1109]

The current transformer unit is a good choice if it offers a manual reset. If you use it to cut the supply when too much current flows, it will immediately sense no current and reconnect again in a continuous loop.

Yeah, and it will same me some time designing and testing a custom circuit. But I thought of an idea to add some additional features. Let me just arrange the idea and put it here.

I think ideally you need something that senses over-current and shuts the power off until the problem is solved and it can be manually turned on again.

Yeah, this will be the main function of the protection circuit I'm aiming for.

The op-amp schematic will work but isn't easily adapted to monitor negative and positive supplies

Yeah I'm still searching for a good circuit.

and I suspect your overload could occur on either or both. It also suffers the cyclical tripping problem.

Brian.

What is "cyclical tripping" ?

--- Updated Feb 26, 2023 ---

Like this one or the earlier one I posted in #13:

If I wanted to monitor all the 4 lines of +/-10V and +/-15V, then I will need 4 of these, it will be very difficult to mount all of them.


I'm thinking to design one circuit which monitor all the 4 lines, and use 4 LEDs with their respected tags to each line. So when a certain line got a short circuit, then the whole system will trip and the relative LED will light up and a buzzer should ring. I'm working on this idea now on proteus.

 

[betwixt]

By cyclical tripping I mean the repeating 'overload/turn off/no overload/turn on again'.

Assuming the overload condition pulls the supply line low, you could do this with a quad open-collector comparator and suitable resistors to scale and offset the voltages at four inputs. Connecting the comparator outputs together (wire ORing them) you get a single signal that tells you if any one or more voltages has gone out of limit. You can use the same circuit for positive and negative rails by monitoring the junction of potential dividers sitting between them with values chosen so the junction should always be positive.

Brian

 

[eagle1109]

Do you exactly know which experiment setup (mis-wiring, overload or short circuit) caused the shown damage?

Yeah, they are the experiments which use DC voltage. It's difficult for me now to remember since I haven't worked in this course for like a year or more.

But I know that the trainees, short anything to the ground, or connect a point that has a high value voltage to a low voltage area.

The IC voltage regulators are internally protected against overload and short circuit

Yeah, but I got multiple ones broken and has marks of burnt color, this one for example:

In this case, the regulator was broken. I think excessive short circuit will break the IC.

but not against externally applied voltage.

Yes, if I'm understanding you correctly, the applied voltage is +/-15V constant.

Unfortunately there are many ways to do so in a wrongly setup power electronics experiment.

Yes, that's the problem, and here comes the story of all the issues with this board.

Well, each semester, the department get certain amount of trainees signing for this course, the recommended number for each class should be not more than 16 trainees for the size of the laboratory having 8 benches as 2 trainees for each bench.

When I started working in this course, I did some mistakes with my basic experience teaching this course. Then after couple semesters I developed a procedure for the trainee to do the experiment. I also got the official lab manual from the company and started taking some experiments and put it in a smaller and arranged one for trainees.

The procedure is that the trainee must do the following:
1. Draw the experiment schematic.
2. Wire the experiment according to the manual drawings.
3. After all trainees are done, I check their wiring, then I tell them to turn on the boards with the oscilloscope.
4. At this point, the trainee should do his work by taking the targeted signals which are mostly, the input and output signals.
And some times, the triggering signals. I explain on the projector where to connect oscilloscope +ve and GND.
5. The trainee must draw the output signals, and do some calculations if required.

I know the problem comes in point #4, because I don't go to each bench and tell each trainee how to connect the oscilloscope +ve and GND so here where it comes the cases of faults in doing the experiment; like, some trainees misconnect something in the experiment wiring and some misconnect the oscilloscope probes.

Another important point, which is that there are no oscilloscope special probes for the holes' size of the board, they way to use it is to take like this oscilloscope alligator cable:

And then, we connect two extending cables from the alligators to the board testing points, but the problem that I believe is a real problem, that the exposed points are a source of shorting the potentials from the test points of the board. Like this:

The black heat shrink bit is my solution to this suspected problem, and I put these protection stuff in almost all the cables in the lab, but when the course transferred to other trainers I came and found that most of them are removed. But I will insist on this issue with the head of the department or use super glue so no one can remove them whatsoever.

Before designing individual protection means I would sketch the expected overload/ damage cases and consider how they can be prevented, respectively if at all.

Yes, but more investigation is a bit difficult for me right now as I'm not working in this course for like couple years and I don't think the department will assigned it to me next semester.

 

[eagle1109]

By cyclical tripping I mean the repeating 'overload/turn off/no overload/turn on again'.

Assuming the overload condition pulls the supply line low, you could do this with a quad open-collector comparator and suitable resistors to scale and offset the voltages at four inputs. Connecting the comparator outputs together (wire ORing them) you get a single signal that tells you if any one or more voltages has gone out of limit. You can use the same circuit for positive and negative rails by monitoring the junction of potential dividers sitting between them with values chosen so the junction should always be positive.

Brian

I'm working on this design, I still have an issue to the op-amp output, and the flip-flop SET/RESET conditions.

Another important thing is that I have to use an external power source to power up the op-amp circuit. But if after all this circuit will protect the board from getting things fried, then it won't be a problem.

 

[betwixt]

The design principle is good but you can make it more versatile by using the other parts of the LM339 to monitor the other supply rails at the same time. If you use a potential divider BETWEEN the supply rails rather than to ground you can make the comparator sense out of range voltages on all the rails while still keeping the measurement point above ground potential. Just make sure the resistor values are chosen so the negative supplies still produce a positive voltage at the resistors junction. Also remember that you can swap the + and - inputs of the comparator to reverse it's output state for a given input.

I would be careful with BUZ1, if you put it in the transistor emitter it will have to pass the relay coil current and it will also increase the voltage needed to make the transistor conduct. It would be better placed across the relay coil with the emitter grounded.

Brian.