design current limiter box for repairs? (light bulb limiter discussion thread)

[coppercone2]

So there is a thread here about a sorensen supply that suggested current limiting to the input of AC powered equipment to aid in repairs. The thread suggested a light bulb.

I thought that maybe this could be taken a step further, so that a rheostat can be wired in circuit.

For people that use this 'light bulb' current limiting trick, do you think a rheostat would work, or does the temperature coefficient of the lightbulb help? What is a reasonable value of resistance range for mains devices in order to pick rheostat?

I essentially wanted to make a box with two connectors so I can plug it into mains between the device under repair.

[TheMG]

A light bulb is better than a resistor/rheostat, because of its non-linear resistance characteristics. When the bulb filament is cold, it has a lower resistance, and the resistance increases as the filament heats up. The halogen variety exhibits this characteristic even more than regular incandescent bulbs.

What this translates to, is less voltage drop at lower currents than a resistor set up for the same current limit.

In my opinion though, such a method of limiting current is not the best when in comes to SMPS troubleshooting. It can cause problems, because most SMPS are designed not to operate below a certain voltages, so you can end up on a wild goose chase while the entire time it was the voltage drop from the current limiter causing the power supply to misbehave.

[coppercone2]

What is the suggested way then? I get that it needs to be used carefully.

Current limiting inverter of some kind?

[David_AVD]

I have a series lamp box I made up years ago and it's very useful, but not appropriate for every situation.

Mine has 2 lamps and 2 switches. One switch adds the second lamp in parallel to the first lamp.

The other switch bypasses the lamp(s) so you can jump to full power without removing the test box from the line.

I use it regularly with SMPS boards, in one or two lamp mode depending on the SMPS size.

It won't protect the DUT in every situation and it may cause instability depending on the DUT's inrush current and circuit design.

[bdunham7]

I thought that maybe this could be taken a step further, so that a rheostat can be wired in circuit.

No, a rheostat won't work.  The light bulb has a very low resistance but will very quickly increase if the current is beyond a threshold.  Sort of like a fuse that doesn't blow all the way.  I guess it really is a PTC!

For some applications, a variac is a way to ramp up the power slowly to spot any problems, but that is a bad idea with many modern devices.  It is even a bad idea for some pretty old devices, like solid-state audio amplifiers with soft-start circuits.  For these, the bulb is still a valid concept, but you have to look at it as a short-circuit-explosion-preventer and use a bulb that won't trip easily even with a bit of inrush.  So you might need a pretty big bulb.  A circuit breaker or easily replaceable cheap fuse is just as valid a solution, IMO.  With these sorts of devices, limiting the current can be counterproductive--you either need to give them full juice or else  fold back (bulb) or shut off the current (breaker).

[edpalmer42]

As mentioned above, there is no perfect solution.  These devices are typically used for a first power up of an unknown device rather than an extended power run, although some are usable for an extended run if troubleshooting an intermitent fault.  Based on my experience, I'd class them as follows:

Series resistor or rheostat:  Poor solution.  Voltage applied to DUT changes with current drawn.  Minimal protection in case of a fault in the DUT because you can't use a large enough value to limit the current without affecting the DUT.

Variac (variable autotransformer):  Excellent solution for a DUT that has a linear power supply.  Good for slow startup where capacitors need to be reformed or a known fault will draw too much current.  Not recommended for SMPS due to unknown response of SMPS to low input voltage and possiblity of supply suddenly turning full on when the voltage reaches a certain point.  This will result in full current draw (perhaps more due to low voltage) and defeats the purpose of a current limiter.

Dim Bulb Tester:  Best solution for SMPS.  Delivers almost mains voltage to DUT while providing current limiting.  I use and recommend a lamp socket and a set of incandescent or halogen bulbs.  You could build everything into a box, but it will be bulky.  Start with a small bulb and work your way up.  My collection goes from 7.5W (~65ma) to 100W (~1.1A).  When I started up a 33V@33A power supply at no load, I started with the 7.5W bulb and, of course, it couldn't start, but nothing smoked either.  I think I had to go to the 100W bulb before it started properly, but it did start properly and safely.

Sometimes, a Variac and a dim bulb tester can be a good combination.  Knowledge of the DUT and its internals would be helpful.  An isolation transformer doesn't fit in this discussion, but may be a useful addition, particularly with a SMPS.

Ed

[coppercone2]

Are there suitable PTC's that can be used for this ? I guess a light bulb has a higher reliability anyway, I just imagined maybe you can get a big rotary switch and solder some PTC to the back to make it somewhat compact and universal. Big clunky rotary switch with PTC soldered to the back of it would be a quick project compared to light bulb holders etc.

I think that the response time of a light bulb will beat out any conventional PTC because of the thin nature of the wiring, so its a better limiter? (the light bulb seems like an almost instantaneous device because of its negligible thermal mass).

Would there be any benefit to putting a small rheostat in series with the light bulb to make some kind of fine trim adjustment.. then you can gradually reduce resistance until it starts up without having to keep so many lightbulbs around? I thought to put the lid for the box on a hinge with a bolt-knob to make swapping bulbs easier, but it would still seem greatly in my favor to reduce the number of light bulbs required. Are there some kind of special vacuum thermal resistance elements that are made specifically for current limiting rather then 'pleasant' light outputs out there? Like optimized wire geometry or something..


How are reactor inductors for this purpose also? A series connected variac or a tunable slug


And I am starting to think getting a light bulb desk lamp or floor pedestal lamp and adding some sockets and a transfer switch on it for cables might be a good option too to keep things practical. secret hacker mods


So in a concise way:
1) special light bulbs? (halogen is one of them, but I mean purpose made stuff.. hot wire hydrogen resistor comes to mind)
2) using PTCs? (small form factor, easy assembly on rotary switch)
3) series adjustable reactor? (series variac or adjustable slug)
4) light bulb + rheostat tuner? (reduce light bulb/parts replacement)
5) best for decour to modify a existing lamp?
6) rheostat + adjustable inductor + light bulb? (this seems interesting).

[David_AVD]

Using a light bulb is good as it gives you visual feedback when it's active. It can also dissipate the heat required. You can also swap out the lamp(s) to suit the situation.

[edpalmer42]

So in a concise way:
1) special light bulbs? (halogen is one of them, but I mean purpose made stuff.. hot wire hydrogen resistor comes to mind)
2) using PTCs? (small form factor, easy assembly on rotary switch)
3) series adjustable reactor? (series variac or adjustable slug)
4) light bulb + rheostat tuner? (reduce light bulb/parts replacement)
5) best for decour to modify a existing lamp?
6) rheostat + adjustable inductor + light bulb? (this seems interesting).

To answer in a concise way:
1) No
2) No
3) No
4) Yes + No
5) No
6) No + No + Yes

Yes, I know I'm evil. 

Okay, that was a little too concise.  To expand on those....

1)  Not necessary.  The magic comes from Tungsten's unique temperature characteristics.  It does exactly what you want for this application so anything special is unnecessary.

2)  A PTC is a replacement for a self-reseting circuit breaker i.e. a circuit interrupter.  That might be a useful item, but not for the purposes of this thread.  A lamp will limit the current without disconnecting it.  That can be helpful in the troubleshooting phase by allowing you to ( carefully!!! ) probe the DUT while it's in an active but current-limited fault state.  This should help you find the fault that would otherwise cause smoke, flames, burned traces - you get the idea.

3)  A series reactor will have a fixed impedance just like a rheostat or series resistor.  That's not the characteristic we want.  We want something that will present a low impedance under normal conditions while switching to a higher impedance, but not an open circuit like a PTC, if the DUT starts drawing too much current.

4)  The rheostat serves no useful purpose here.  See #3.

5)  Not necessary.  Since you're in the U.S.  buy a standard duplex outlet, wall box and cover.  Break off only one of the tabs on the side and wire the two terminals that you've just seperated to a power cord.  You've now put the two outlets in series.  Label the box so you or someone else doesn't mistake it for a home made extension cord.  Plug the DUT into one outlet and a standard lamp into the second outlet.  You can use a desk lamp, trouble light or whatever's available.  I don't know if there's anything similar that's available in the other parts of the world. 

6)  The rheostat & inductor serve no useful purpose here.  See #3.

Ed

[coppercone2]

So essentially the light bulb is fast enough that any series impedance is not necessary at all. I thought raising the line impedance might help but I realize now maybe the light bulb goes into a thermal runaway of sorts so it works better as a limiter when there is a surge and goes into a protective region even faster?

The idea was that the impedance element is 'ever present' and thus 'instantaneous' compared to the light bulb which still is defined by thermal time constants of some sort.

But I suppose if you choose the right light bulb, you can get the 'correct' cold resistance. I don't know how much variance there is in resistance (i.e. how many different light bulbs are too much). I see someone on youtube built a box with 4 light bulbs that can be switched in and out in parallel/series configurations... which is starting to get slightly crazy, hence the desire for a single light bulb and a adjustable impedance element of some sort. That machine he has there is a little big for my taste. 




Given that the machine in the video does exist and someone made it, are you sure there is no room for a rheostat to try to reduce it? I guess part of the problem is that the light bulb controlled proportion of the impedance would be too small in comparison to the whole thing.. but I was hoping maybe to change 4 light bulb design to a equivalent two light bulb design.

And by PTC I mean a inrush limiter PTC not a cutoff, I guess the problem is I am thinking of a NTC not a PTC (switchmode power supply inrush limiter), not sure if PTC come in a 'mild' form? I never really thought about how fast a NTC in a power supply is.

[Shock]

Currently I just use a series light bulb (aka lamp, globe) if required and I have different wattages. My solution when I want to waste more time is to make a permanent series parallel/socket setup with either bypass or selection switches to choose the wattage. Just makes it easier than inserting and removing bulbs.

I have a variac and isolation transformer as well. It's useful knowing the mains supply and output voltages and the output current and power. Installing inrush current limiting and adequate circuit protection is also desirable (if you have more than just a simple dim bulb setup).

Edit:

I think a rheostat could work in a low power situation but they tend to produce heat. My setup is 240V @ 10A so both a large isolation transformer and variac. Certain isolation transformers can be wired to provide step up voltage and/or include voltage taps, so they have several advantages.

Since the dim bulb setup is a reactive load I think it's far better for it's intended purpose alone as short circuit protection. If you wanted to add variable voltage for graceful power ups and dialing in the exact voltage, add a variac.

For doing a lot of 120/240V work, use a step up/down autotransformer. Mine has configurable input and output voltages which are useful in dialing in the correct output voltage for the dut. I typically use this to run 120V gear up for longer burn in periods as it frees up my test setup. Most of my gear though has been converted over to 240V.

All this was fairly inexpensive to setup secondhand.

[edpalmer42]

So essentially the light bulb is fast enough that any series impedance is not necessary at all. I thought raising the line impedance might help but I realize now maybe the light bulb goes into a thermal runaway of sorts so it works better as a limiter when there is a surge and goes into a protective region even faster?

The idea was that the impedance element is 'ever present' and thus 'instantaneous' compared to the light bulb which still is defined by thermal time constants of some sort.

But I suppose if you choose the right light bulb, you can get the 'correct' cold resistance. I don't know how much variance there is in resistance (i.e. how many different light bulbs are too much).

Careful, you're overthinking this.  This isn't a laser-focussed piece of precision test equipment.  It's one step above a crowbar and duct tape!  There is no 'correct' value.  There's just a 'this might be okay' value.

I see someone on youtube built a box with 4 light bulbs that can be switched in and out in parallel/series configurations... which is starting to get slightly crazy, hence the desire for a single light bulb and a adjustable impedance element of some sort. That machine he has there is a little big for my taste. 

I was quite intrigued to note that his 4 bulbs allow him to hit every 20 W increment from 40 W to 200 W except for 180 W.  Very nice, but I don't see the value in that level of precision.  One good thing about his permanently mounted bulbs is that you're not juggling glass bulbs.  Depending on how fumble-fingered you are, you might find yourself cleaning up shattered bulbs. 

Given that the machine in the video does exist and someone made it, are you sure there is no room for a rheostat to try to reduce it?

What value of rheostat would you use?  5 ohms?  No, that would 'limit' the current to 24 Amps!  100 ohms?  That would limit the current to 1.2 A, but the cost would be an unacceptably high voltage drop under normal conditions.  For example, if the DUT is supposed to draw 100 ma, the voltage drop across the 100 ohm resistor would be 10V.  That might be acceptable, but you're approaching the limit.  A higher current DUT will make things worse.  And let's not forget about the power dissipation.  Your 100 ohm resistor will have to be rated for 144 W.  150W rheostats are certainly available, but they're a bit large - about 4 inches in diameter.  They can also be a bit expensive, although I see some cheap ones coming out of China.

I guess part of the problem is that the light bulb controlled proportion of the impedance would be too small in comparison to the whole thing.. but I was hoping maybe to change 4 light bulb design to a equivalent two light bulb design.

Well, if you did something similar to the video, but with just 40W and 60W bulbs, you'd have the equivalent of a 40, 60, and 100 W limiter.  That would be suitable for larger devices but not for smaller ones.  I'm not sure if that's a deficiency or not.  Smaller devices might actually be more difficult to protect due to their generally lower power devices that will overload and blow at low current or power.

And by PTC I mean a inrush limiter PTC not a cutoff, I guess the problem is I am thinking of a NTC not a PTC (switchmode power supply inrush limiter), not sure if PTC come in a 'mild' form? I never really thought about how fast a NTC in a power supply is.

An NTC thermistor starts at a high resistance and, as current flows through it, it heats up and the resistance drops.  This allows the input capacitors to charge while limiting the current spike that would otherwise be present.  That's the opposite to what we want.  I don't know if a 'mild' PTC is available but, again, the question of power dissipation comes to mind.  Also, both types of device would be relatively slow to 'reset' to their initial states so you'd have to wait a bit between tests.  I really have no idea if that would be a practical issue or not.

Ed

[David_AVD]

The OP is indeed overthinking this. Just make a box with a lamp or two, switches and start using it. Make a basic one now and improve on it later if you really need to.

The series lamp won't protect the DUT in every situation. Sometimes things can still burn up without drawing enough current to light the lamp fully.

There's no substitute for paying attention to what the lamp and the DUT are doing.

Also keep in mind that some DUT adjustments may be invalid if only done / checked while the lamp is in series due to the reduced voltage.

[coppercone2]

what I need is a vacuum tungsten rheostat to leave it at one bulb lol

And I generally make my design out of plywood with metal grate vents and it comes together pretty nice (I try to make things look kind of like general radio old wood equipment) so resizing/modding is usually a big waste of materials (I stain, varnish and polish my equipment boxes), I just wanted to get it right the first time.

I think I found one old ceramic base for a light bulb, I will just maybe leave it primitive until I get some experience working on it to see if it needs more.

[coppercone2]

I have two ceramic light bulb bases in the junk box, so I guess my limiter will have two bulbs

[coppercone2]

I don't know much about light bulb series, what should the space allowed be for light bulbs inside a DBL box?

Is there any benefit towards accommodating a large bulb like a heat bulb?

How many watts should I shoot for the maximum limiter bulb size to be.. does anyone say use 500W bulbs in theirs? Those are 8 inches tall...

What is minimum safe distance between 500W bulbs and chassis and other bulbs? I think a reasonable maximum limit would be 2x 250W bulbs.. and load it with like 50 or 100W normally. I don't think there is much reason to build it to accommodate 500W bulbs.. or is there? I usually build my chassis so the top comes off... so in the rare case I need that much power I can run it without the hood essentially.

I guess I need to get a incandescent and then put the enclosure material nearby and measure temperature to find out, I wanted to use perforated sheet of steel.

[Audiorepair]

Such lightbulbs are designed to show a fault.

Once that fault is determined, the power will be shut off by the person conducting the experiment.

I think you are are somewhat over complicating things here.

[Shock]

Having multiple sockets adds flexibility to make it easily adjustable. Have larger bulbs if you need them but I'd aim to start with a couple each up to 100W.

The wattage used is based on the load and how much you want to limit it, you can find information on wattage here: https://antiqueradio.org/dimbulb.htm.

[coppercone2]

Such lightbulbs are designed to show a fault.

Once that fault is determined, the power will be shut off by the person conducting the experiment.

I think you are are somewhat over complicating things here.

I don't like building things that don't cover all the angles. The carpentry alone makes it worth it to do pretty much anything electrically. And since I am putting banana plugs on it, it can also work as a dummy load, so indefinite on-time is useful.

[Shock]

If you want audio dummy loads, lowes.com sells item #362342 the Utilitech 240V Universal (water heating) element. They are 8 ohms rated up to 3500W if cooled.

[m3vuv]

dont know about other places but as far as i know retailers here in uk can only sell there current/old stock of incandescent bulbs here,getting hard to find,bit like non moulded 13amp plugs.

[Shock]

You can still buy incandescent bulbs look for halogen, fancy, party, industrial bulbs.

[m3vuv]

i was talking about normal household light bulbs.

[Shock]

Which is why I said about those other sources.

[edpalmer42]

Here, you can still find incandescent bulbs at various dollar stores.  I don't know if it's a standard product or if they're just using up old stock.

You can also find 'rough service' bulbs that are sometimes coated to protect the glass, have a more rugged filament, and are rated for 130V.  Come to think of it, if you're juggling bulbs, they might be the perfect ones to use.