Check the electronic board, with lamp and variac

[CharlotteSwiss]

When I have a board failure, perhaps to find a short circuit, or after replacing components, I usually power the device with a lamp in series, gradually increasing the voltage with a variac until I reach 230V.
I have a big question: could powering a device with a voltage lower than 230V, gradually, cause damage to some component or part of the circuit? Thanks for your attention.

[thephil]

As far as I know, most circuits will be fine but some SMPS don't like undervoltage at all. They will try to produce the correct output voltage/current at too little input voltage and may get damaged in the process.

So I use the dim bulb but usually no variac for that purpose.

[ME]

it maybe of note that the 110v rail in the psu of the plessey rt-320 txcvr normaly blows up when the battery is getting low.

[CharlotteSwiss]

So if I find the small transformers with yellow tape on the board, is it better to avoid increasing the input voltage with a variac? So almost all boards...  But am I wrong or could in this case it start from 100 V AC, and then go up to 230? Because if I remember correctly, those power supplies have that operating voltage range (?). Thanks

[pcprogrammer]

There are many variables to consider here, but when it comes to SMPS's a lot of them will work from 90V to 250V or even from as low as 65V to 265V. What one should keep in mind, is that at some nominal load at the output of the SMPS the input current will rise with the lowering of the voltage, because the amount of power drawn still has to match the amount of power delivered plus the losses in the SMPS.

This means that the variac used to lower the mains voltage has to be able to provide this power.

Having a lamp in series does reduce certain risk, but in the case of a short in the device under test there will still be the current the lamp is rated for flowing through the system. So the lamp has to be chosen accordingly. I wonder what a led lamp will do in this case. 

[CharlotteSwiss]

So, to power a DUT with a switching power supply, even if you want to use a Variac, it would be best to start with at least 100V AC, and not less?
Regarding the current flowing through it, my Variac has a 15A fuse, but I power it with a 10A circuit breaker. For me, the important thing would be to avoid further damage to the board under test, which is why I asked this question, as I had some doubts about it. Thanks.

[pcprogrammer]

So, to power a DUT with a switching power supply, even if you want to use a Variac, it would be best to start with at least 100V AC, and not less?
Regarding the current flowing through it, my Variac has a 15A fuse, but I power it with a 10A circuit breaker. For me, the important thing would be to avoid further damage to the board under test, which is why I asked this question, as I had some doubts about it. Thanks.

If the switching power supply is rated for working of a low voltage sure. One thing to keep in mind that with a fully operational switching power supply the output voltage will be the nominal one and it will not be lower due to the input voltage being lower. In case of a normal old style transformer based power supply the output voltage will be lower and the system behind it might not like that.

So what to use in a test case is always based on the specific setup. There is no fit for all one solution to this question.

Also make sure to have a RCD (ground fault protection also called earth leakage circuit breaker or disjoncteur différentiel in French) in the mains setup to protect yourself when working with mains voltage.

And be aware that poking around in a SMPS the mains voltage is rectified to a high DC voltage that is more lethal than the AC mains. In for instance the SMPS's in desktop computers one will find nice big capacitors rated for 400V DC that may still hold some charge and pack a mean bite when touched on the connections. The cooling fins on the primary side might also carry the high voltage, so be very careful when working on these things.

[CharlotteSwiss]

Thanks pcProgrammer. So from what I've learned, if you need to run tests on a device powered by mains voltage, which also includes a power supply, it's best not to use a variac when troubleshooting, but to power it directly with mains voltage. It might, however, make sense to use a variac for devices that operate without a built-in power supply (for example, an electric heater, etc.). For safety, certainly. I always work with RCD protection, and I'm very careful with input capacitors (320dC), which I discharge before taking measurements when I unplug the mains plug. Some heat sinks on the primary circuit are also dangerous; I've already had the opportunity to verify this. Thanks

[CharlotteSwiss]

I still have a couple of doubts about using the variac and the bulb in series, while I check that there are no short circuits or faults 

I read that powering a board with a variac (even at standard line voltage) while simultaneously taking measurements with an oscilloscope could damage the instrument. Is this true? 

How many watts should the bulb be in series? I use a 60W bulb...

Thanks

[SteveThackery]

I'm about to overhaul several old valve amps. It is useful to be able to gradually increase the supply voltage. Of course there are no switch-mode converters or voltage regulators, so nothing that could be damaged by a low supply voltage.

[CharlotteSwiss]

Artificial intelligence suggested that it would be best for valves not to start with low voltages. I also knew that variacs were very useful for valves. Maybe artificial intelligence lives on Mars...

Anyway, as for when to use the variac, let's say I've figured it out. I still have the two questions from my previous post 

[thephil]

I read that powering a board with a variac (even at standard line voltage) while simultaneously taking measurements with an oscilloscope could damage the instrument. Is this true? 

No – I can't see how that would happen. I believe those warnings may have been referring to isolation transformers not variacs. Of course it is perfectly fine to take measurements using those, too. However, it is very easy to create very "interesting" situations when probing around with a normal (grounded) oscilloscope in a circuit that has been isolated. I.e. things may no longer be as isolated as they seem, resulting in damage to you, the oscilloscope or the DUT (Dave has a very good video on that topic). But I cannot think of anything particularly problematic when using a regular variac (which is not isolating)

How many watts should the bulb be in series? I use a 60W bulb...

Depends on the device you are testing. The bulb determines the maximum power applied to both, so if your device is shorted, it will conduct 60W/230V = 260mA. Make a guestimate if that will explode/melt it or not. Just keep in mind, that the device under test needs to be able to draw enough current to actually work. That has confused me a couple of times ...

[CharlotteSwiss]

I read that powering a board with a variac (even at standard line voltage) while simultaneously taking measurements with an oscilloscope could damage the instrument. Is this true? 

No – I can't see how that would happen. I believe those warnings may have been referring to isolation transformers not variacs. Of course it is perfectly fine to take measurements using those, too. However, it is very easy to create very "interesting" situations when probing around with a normal (grounded) oscilloscope in a circuit that has been isolated. I.e. things may no longer be as isolated as they seem, resulting in damage to you, the oscilloscope or the DUT (Dave has a very good video on that topic). But I cannot think of anything particularly problematic when using a regular variac (which is not isolating)

How many watts should the bulb be in series? I use a 60W bulb...

Depends on the device you are testing. The bulb determines the maximum power applied to both, so if your device is shorted, it will conduct 60W/230V = 260mA. Make a guestimate if that will explode/melt it or not. Just keep in mind, that the device under test needs to be able to draw enough current to actually work. That has confused me a couple of times ...

Thank you, very kind of you.

Okay, so there's no problem using the oscilloscope when I use the variac (I thought so too, only I read some misleading information...).

On the other hand, the discussion about the power of the bulb in series has left me quite confused. I know that the bulb turns on (perhaps even at low intensity) ONLY if there's a short circuit on the board it's powered. Otherwise, it stays off.
But if, for example, I'm using a 60W bulb, does this mean it can only draw 230mA from that short circuit?
I've always used the bulb, but honestly, given its power, I don't understand what currents will be in play on the board I'm testing

[nonius_]

To the best of my knowledge I've never damaged equipment by powering it with a series-lamp limiter (dim bulb) and a variac.

But I'm sure there could be special cases where damage might occur. Then again, I'd wonder how well-designed the equipment is in that case. Because, if for example, there'd be a high-impedance connection in the grid (power plug), that should/would cause similar damage.

Anyway, I think you're overly concerned.

And the power-rating of the bulb depends on the DUT (its power rating and the type of equipment) and whether it starts up properly with a relatively low-wattage series lamp. In my dim-bulb tester most of the time there's a 100W bulb. That bulb is of too low rating to properly start the powersupply of my PM3055 (rated at 47W). There appears to be a threshold below which the PSU doesn't work. But most other small equipment, and valve-radios, work fine with that lamp.

Valve-radios are much less critical in that respect, they still work mostly at substiantially lower line-voltage. Perhaps not so loud, or not so sensitive, or...., but still work good enough to verify there are no major problems.

Conclusion: it varies. Experience and common-sense are required when using a series-lamp limiter. The more you use one, the better you get at interpreting it.

[MarkF]

. . .
Having a lamp in series does reduce certain risk, but in the case of a short in the device under test there will still be the current the lamp is rated for flowing through the system. So the lamp has to be chosen accordingly. I wonder what a led lamp will do in this case. 

The whole premise behind the 'dim bulb' tester is the properties of the filament.
As the filament gets hot, the resistance goes up. 
Which means the current goes down protecting you from fire.

You have to make a choice of the bulb wattage (current limiting) to protect the DUT.
Remembering that the brighter the bulb gets the more current limiting AND the less voltage the DUT gets.
(Example:  If the DUT is a second bulb the same wattage as in the tester, the voltage across each bulb will
be half and the current will be half.)  You have to give up some current limiting so that the voltage to
the DUT is not impacted too much.  A delicate balance.

I have two bulbs in parallel in my tester to provide a more flexible choice of current limiting.

This has gone a little off the rails reguarding the OP's original question.

LED lamp and fluorescent bulbs are a big NO NO!

[CharlotteSwiss]

Thanks for the advice.
I'm having trouble understanding the current quantities involved when using the light bulb in series with the device being tested. For example, I don't understand what current level a 60W bulb will start to light up (in the event of a short circuit on the board being tested). The only thing I know for sure is that I've always used filament bulbs (never LEDs or fluorescents).

[pcprogrammer]

The whole premise behind the 'dim bulb' tester is the properties of the filament.
As the filament gets hot, the resistance goes up. 
Which means the current goes down protecting you from fire.

This is exactly the point of the "dim bulb" tester. As long as the current stays low enough to not bring the needed heat to the filament, the resistance stays low and the limiting of the voltage on the DUT is small enough to not impact its working. As soon as the current goes up to high the filament starts to glow and the resistance of it will increase.

You can check the cold resistance of a bulb with a multimeter and use the result to calculate the initial current. The argument of thephil that the DUT in case of a short only conducts 60W/230V = 260mA is only after the bulb has lit up. There will be a higher current at first.

My wondering about the LED lamp is broader than this property of an incandescent lamp, because it also rectifies the voltage. For an SMPS based device not too big of a problem, but for a transformer it will be a different story. I also added the question to bring awareness to the fact that you can't just use any lamp for this. So yes to the statement of MarkF.

LED lamp and fluorescent bulbs are a big NO NO!

On the subject of probing the DUT with an oscilloscope it always needs proper precautions when mains is involved. Before probing make sure your expectations of what you are going to measure are correct. Think about peak voltages and how much your scope input is rated for in respect to the probe you are using. Best to get a proper non switchable 10:1 or even 100:1 probe for this. Make sure the scope is either not grounded and that the live signal is not connected to the case (This still is no guarantee the case will not be at some potential though) or use an isolation transformer while keeping the scope grounded. In this case there are still risks of shock of course.

The best solution is a differential probe rated for the high voltage in question, but for higher frequencies these are expensive.

My advice is to always analyze what you try to do first and asses the risks involved before going at it.

[CharlotteSwiss]

Now the operating principle of the series lamp is a little clearer.
If there is no short circuit on the board, the current flows along the cold filament of the lamp (there will only be a low resistance). When the current required by the board increases (due to a short circuit), it turns on the lamp, the filament increases its resistance, and limits the current to the board. If during a short circuit, for example, we use a 60W lamp, the maximum current flowing to the board will be limited to 230mA (assuming the lamp is fully lit). In this situation, if the board being tested were operating at a higher current, it would not even be able to function properly due to the lack of current.

As for the oscilloscope, I have the differential probe. But my question about the safe use of the oscilloscope with the variac was related to the use of traditional passive probes. As far as I know, the oscilloscope's ground clip is referenced to the mains ground, so when I use the oscilloscope, I must always connect it to a 230V socket, which has the pole connected to the house ground.

[pcprogrammer]

Now the operating principle of the series lamp is a little clearer.
If there is no short circuit on the board, the current flows along the cold filament of the lamp (there will only be a low resistance). When the current required by the board increases (due to a short circuit), it turns on the lamp, the filament increases its resistance, and limits the current to the board. If during a short circuit, for example, we use a 60W lamp, the maximum current flowing to the board will be limited to 230mA (assuming the lamp is fully lit). In this situation, if the board being tested were operating at a higher current, it would not even be able to function properly due to the lack of current.

Correct, and that is why one has to match the wattage of the lamp to the target.

You mentioned an electric heater earlier, and if that needs lets say 240W to function properly it takes 1A of current. With a 100W lamp in series it will never get that current and warm up less. The voltage will split between the lamp and the space heater based on their resistances, which depends on the temperature of the lamp filament and in some cases also that of the heater element.

As for the oscilloscope, I have the differential probe. But my question about the safe use of the oscilloscope with the variac was related to the use of traditional passive probes. As far as I know, the oscilloscope's ground clip is referenced to the mains ground, so when I use the oscilloscope, I must always connect it to a 230V socket, which has the pole connected to the house ground.

There is a caveat here. If there is a potential difference between the neutral and the ground it can trip the RCD when connecting the probe ground lead to the neutral. This happens to be the case where I live. The neutral is coming from the pole transformer several hundred meters away and is not connected to ground in my house. It is grounded near the pole transformer though. Due to this there is a potential difference between my own ground and the neutral and the RCD trips when I connect them together.

Has partially to do with capacitive coupling.

For mains measurement I would have to get myself a differential probe, but I have not had the need for it so far. I tend to stick to the secondary side of power supplies. 

[SteveThackery]

Okay, so there's no problem using the oscilloscope when I use the variac (I thought so too, only I read some misleading information...).

I really don't think you should ever say that. The important thing is to watch Dave's video so that you have a complete understanding of why and when problems can arise, and how to avoid them. You need to understand what is safe and what isn't.

[CharlotteSwiss]

Let's take a step forward and see if I understand this damned series lamp: when I don't have any short circuits on the board under test, the lamp remains off, meaning the filament is cold and its resistance is close to zero. In this case, if the board requires 1A, the current flows easily through the lamp.
Let's say I have a short circuit on the board, requiring 1.5A: if I have a 60W lamp, it turns on and lets a maximum of 1.3A through, limiting the current to the board (though we'll still have an excess of 0.3A on the board). If I were to use a 100W lamp, it turns on and lets a maximum of 1A through (the board's nominal voltage, so better).
Have I taken a step forward, or two steps back? 
Thanks

[pcprogrammer]

Let's take a step forward and see if I understand this damned series lamp: when I don't have any short circuits on the board under test, the lamp remains off, meaning the filament is cold and its resistance is close to zero. In this case, if the board requires 1A, the current flows easily through the lamp.
Let's say I have a short circuit on the board, requiring 1.5A: if I have a 60W lamp, it turns on and lets a maximum of 1.3A through, limiting the current to the board (though we'll still have an excess of 0.3A on the board). If I were to use a 100W lamp, it turns on and lets a maximum of 1A through (the board's nominal voltage, so better).
Have I taken a step forward, or two steps back? 
Thanks

The first part is kind of correct. The resistance of the lamp is low enough but it depends on the wattage of the bulb. I just measured two 40W bulbs and they showed about 95 Ohm, so the startup current is ~2.5A. The nominal current is only 0.1666A which means a resistance of ~1440 Ohm. (This based on 240V)

As long as the target board draws significantly less than that 2.5A the lamp resistance will hardly change and the voltage drop over the lamp will be low enough. But it very much depends on the device under test. In case of a larger inrush current the device might never get to a normal point of operation, because the lamp will be limiting both the current and the voltage.

The simple theory is two series resistors that as such make up a voltage divider. Both resistors are in play as to what the current is going to be and who gets the biggest part of the voltage supplied. With one or both of the resistors having a large dependency on the temperature the result will vary over this temperature change.

For the second part, if there is something wrong with the board and it takes a higher current than normal, based on your premise 1.5A, the lamp will initially take that 1.5A and start to warm up and increase its resistance. This will cause the current to drop. The device might then not act as a short anymore and let the current drop a lot more, cooling down the lamp. This leads to a voltage rise on the device and it might act up again, etc.

It is a dynamic system at play here. Based on the lamps wattage it is possible to calculate the nominal current of the lamp. In case of a real short in the device under test, it is this current that will be flowing once the lamp is fully on. The benefit of the lamp in series is that in case of an actual short, the initial current is limited to the startup current of the lamp and not a high enough current that can evaporate circuit board traces.

This is a problem with fuses. Even though a fuse for instance is rated for 1A, it does not limit the current to 1A. It takes time to blow a fuse, which depends on the type of fuse. Slow or fast are two types, and putting a fast fuse in a device with a high inrush current will blow it every time the device is powered on. In the case of a fuse and also the lamp, the max current depends on the total resistance of the circuit from the source to the device. Wiring has resistance to.

Hope this clears things up a bit for you.

[CharlotteSwiss]

Thanks for your very detailed explanation. I'll think about it carefully over the next few days, with my lamp kit at hand, and I'll let you know if everything is clear. But I already feel like I've made two steps forward! 

[thephil]

Here is what I did when I first learned about the dim bulb tester and did not understand how it worked exactly and what wattage to pick: I put a bulb behind my variac and various other things (including a piece of copper wire) in series and started measuring voltages and currents at different input voltage settings and DUT resistances. Nothing is more enlightening than a few experiments of your own.

[CharlotteSwiss]

Here is what I did when I first learned about the dim bulb tester and did not understand how it worked exactly and what wattage to pick: I put a bulb behind my variac and various other things (including a piece of copper wire) in series and started measuring voltages and currents at different input voltage settings and DUT resistances. Nothing is more enlightening than a few experiments of your own.

That's what I'll do next week, then I'll write my experience here. Thanks