4
\$\begingroup\$

For a small 24V 1A transformer where should I place the fuse? Secondary 1A fuse or primary 100mA fuse (or above accounting for inrush current)? Mains is 230V. Also, the smallest fuse available in the market seems to be 200mA.

Also, how about a 30V 6A transformer? I guess for a 12V 500mA transformer the answer is obvious, as I have no access to 50mA fuses for the primary. But how useful is a primary side fuse? It's not like the transformer will short out on its own, and it's not like the utility voltage will go up to extremely high levels for any length of time. So I am really placing the fuse to protect the transformer from a shorted load placed on the secondary (including shorted bridge rectifier), right?

Added: Thanks for all the replies. These transformers are for one off projects. They are all EI core. So a primary side fuse with 150% rating at 230V would be:

transformer    required fuse    actual fuse
24V 1A         156mA            200mA
30V 6A         1.17A            2A
12V 500mA      26mA             200mA

Is that correct? The minimum fuse I can get is 200mA. And above 1A is usually in 1A steps. I might not be able to get a 1.5A fuse. Also, I have no idea if these are slow blow fuses. The cartridge fuses seem to be tinned copper in a glass tube. I could potentially test the fusing current and time for the fuses just to be sure (fuses are cheap).

How much current might a transformer draw at startup? I found somewhere (lost the link) that it may be a lot more (like upto 35A for a 5A transformer, so it is a bad idea to repeatedly switch on and off a transformer e.g. with a contactor).

The small 12V transformer is possibly inductance limited. I was unable to draw more than 650mA from it at the rated voltage (12VAC). Is this how a class 2 transformer should behave? Or should 500mA have been the short circuit current?

If I understand correctly, a primary fuse will protect the transformer from shorts in the winding. Shorts in the winding can arise due to overheating. Overheating can occur due to high current draw in the primary, or inadequate cooling. There are 2 reasons for high current draw in the primary: 1. repetitive inrush current due to repetitive switching on and off the primary side, and 2. high current draw on the secondary side. If I operate the transformer only with a manual toggle switch (not a doorbell), and if I have a suitable fuse on the secondary, then I have none of those problems. That leaves only inadequate cooling which can destroy the transformer winding. Is my thinking correct?

Is it correct that the main reason to place a fuse on the primary is that a fuse on the secondary is no more necessary to protect the transformer? A shorted secondary will draw enough current to blow the primary side fuse. But this will not happen for a class 2 transformer. So a class 2 transformer might still burn up because it was operating for an extended time at slightly over maximum rating, if it lacked a secondary fuse.

[Sorry for so many questions. I am definitely confused about this. My gut says to just go with 2 fuses and avoid headache... but on some sites they even say 2 secondary fuses!]

added Thanks again for all the replies and clarifications. I found in my hands yesterday a brand new 30V 6A transformer. As soon as I plugged it in (no secondary load), the fuse on the primary side blew. It was rated 5A. Thinking it may be an old fuse, I replaced it with a new one. It blew again. The transformer as well as the mains wire to it was already warm. So I guess that answers my own question. Now for the hassle of returning the transformer...

\$\endgroup\$
  • 4
    \$\begingroup\$ Primary! Consider if the fault is in the transformer itself! \$\endgroup\$ – winny Apr 10 '18 at 10:57
  • \$\begingroup\$ Is there any reason for expecting a fault in the transformer itself, if the secondary is adequately protected against excess current draw, and if I monitor the core temperature? \$\endgroup\$ – Indraneel Apr 11 '18 at 2:58
  • 1
    \$\begingroup\$ @Indraneel Same reason for expecting a fault in any other component: stuff goes wrong. \$\endgroup\$ – immibis Apr 11 '18 at 4:35
  • \$\begingroup\$ Also safety agency/government requirement. \$\endgroup\$ – winny Apr 11 '18 at 5:40
  • \$\begingroup\$ Manufacturer defects would be a reason for a fault in the transformer itself! \$\endgroup\$ – Doktor J Apr 11 '18 at 16:15
3
\$\begingroup\$

Not using any primary fuse could lead to catastrophic failure in the transformer itself - transformer wires are typically insulated only by very thin and brittle lacquer, and a winding shorting on itself due to damage to that insulation will effectively act as a shorted low-voltage, ultra high current secondary, and heat up significantly. Even worse, the lacquer insulation is not very resistant to extreme thermal conditions - many types of transformer wire can (and sometimes are intended to) be stripped of their insulation by a very hot (say, 380-450°C) soldering iron. Also, depending on the core material used, a serious failure could heat the core above the curie point (though that is unlikely with 50/60Hz transformers and their iron cores), completely changing the magnetic properties and causing further escalation in some circuit variations. End result of all these would be a cascading, catastrophic thermal failure.

Why not both - a primary fuse can be optimized to be permissive of any normal operating condition of the transformer, while one or more secondary fuses can make sure the load to the transformer is not faulty - especially when there are multiple secondary windings.

If the transformer either only has a single secondary winding, or could withstand any overload on any secondary winding without catastrophic failure, and if downstream circuitry does not need further protection, a primary fuse can be enough, but will need empirical and/or mathematical optimization that takes the transformer's efficiency into account.

A lot of designs seem to rely on the overcurrent protection features of linear or switching voltage regulators used directly behind the rectifier+filter - which might be OK if you can rely on either the transformer's source resistance or the primary fuse to limit the current in case any of these components fails short.

\$\endgroup\$
  • \$\begingroup\$ My gut does say both, but that does not seem to be the common practice. I want to know why. \$\endgroup\$ – Indraneel Apr 11 '18 at 4:05
  • \$\begingroup\$ Thanks for the explanations! That was very helpful in understanding the source of faults, and design requirements at the different stages. \$\endgroup\$ – Indraneel Apr 12 '18 at 3:39
7
\$\begingroup\$

You should put it in the primary, to protect against transformer faults. This may be optional if it is a Class 2 impedance protected transformer or has an internal fuse or thermal cutout.

Keep in mind that the output load (especially if it's the usual capacitor filter) and the transformer itself will cause inrush surges at turn-on. The latter can be especially bad with toroidal power transformers sans inrush limiting devices such as NTC thermistors (can be tens of amperes for a small transformer) and occurs as a result of when the power was last removed in the AC cycle and when it is reapplied (perhaps counter-intuitively, zero crossing is the worst case).

Continual surges can lead to early death of the fuse if it is not appropriately rated and of the appropriate type (often slow-blow is required).

The fuse has several purposes- to protect the transformer against a short on the secondary. To protect the transformer against an overload on the secondary. To limit the damage in case the transformer itself fails (think shorted turns and smoking), and (in some cases) to protect the mains cord from excessive current (from a short or overload).

\$\endgroup\$
  • \$\begingroup\$ And fourthly, the fuse protects against a badly manufactured transformer! Thanks for the detailed andwer. Many helpful points. \$\endgroup\$ – Indraneel Apr 12 '18 at 3:24
4
\$\begingroup\$

I would suggest to use the fuse at the primary side. The main reason behind this is that fuses are less sensitive to changes and come into role only when there is a drastic change in current value. Moreover, just keep in mind that the fuse should be selected with a tolerance of 150% from the limit/maximum value.

For more clarity, just visit this link http://sites.ieee.org/fw-pes/files/2013/01/transfguide.pdf

\$\endgroup\$
  • \$\begingroup\$ Thanks for the link. Will take a while for me to understand it. \$\endgroup\$ – Indraneel Apr 11 '18 at 4:07
  • \$\begingroup\$ Take your time and get back if still you face any issues. I would be glad to help you out again. ;) \$\endgroup\$ – Sumit Panse Apr 11 '18 at 4:23
4
\$\begingroup\$

It's not like the transformer will short out on its own

Don't bet on it.

Why do plugs have fuses in them? Answer - to protect the infrastructure of the home i.e. the house wiring in the walls and therefore prevent a fire. What your fuse does is both protect the house wiring AND, stop the transformer burning should it develop a fault or the load current rise too high.

Put it in the primary circuit.

\$\endgroup\$
  • 2
    \$\begingroup\$ Plugs only have fuses in them in the UK. \$\endgroup\$ – τεκ Apr 10 '18 at 12:17
  • 1
    \$\begingroup\$ @τεκ I don't dispute that. \$\endgroup\$ – Andy aka Apr 10 '18 at 12:21
  • \$\begingroup\$ I can see the primary fuse protecting the house wiring in all cases and the transformer in some cases. To make it bulletproof, use a secondary fuse too. Is that it? \$\endgroup\$ – Indraneel Apr 11 '18 at 4:11
  • 1
    \$\begingroup\$ @Indraneel ask yourself - what is the secondary fuse trying to protect? If it's trying to protect the transformer from overload then this will be also done by the primary fuse. If you are trying to protect something else then maybe. All you have to do is ask yourself what the fuse is trying to do and, given the answers on this page, if you can't think of a good reason, then why add the cost/unreliability of an extra fuse? \$\endgroup\$ – Andy aka Apr 11 '18 at 8:55
  • \$\begingroup\$ Based on my last experience, I'll put the primary fuse in and worry about the secondary fuse later. I guess I'm thinking the secondary fuse may protect against slow overheating due to marginally higher current draw (which will be missed by the primary fuse), but am not too sure if it's the correct way to do it. Maybe the circuit attached to the secondary should be designed with current limiting. \$\endgroup\$ – Indraneel Apr 12 '18 at 3:29
2
\$\begingroup\$

If this is a pre-manufactured "wall wart" or "lump in cord" transformer, then leave it alone. The assembly is UL listed (or equivalent listing agency, CE is not one). The safety of that assembly has been accounted for, and you void its listing by altering it.


If this is your own device in your own chassis, for which you might seek UL listing, then UL publishes documents to say what they expect. Those are good guidelines to follow.

A 1-amp transformer that fails internally won't draw 2 amps . It will draw lots of amps, or will ground-fault, or will arc. These are detected by the house's overcurrent protection (13A, 15, 16 or 20A depending on territory), its GFCI/RCD devices, or its AFCI, respectively.

If this is a one-off device, one answer is simply to fit AFCI and GFCI/RCD protection upstream of the device (and make sure to ground the metal chassis, to make ground faults more detectable, current returning via ground will cause a rapid trip).

On the other hand, a transformer is drawing 200% of spec has too much load on its secondary. It can endure this for a short time, just as car starter motors do.

\$\endgroup\$
  • 1
    \$\begingroup\$ It's not UL listed, and I'm not applying for UL listing. Just want to protect the transformer from burning up. \$\endgroup\$ – Indraneel Apr 11 '18 at 4:31
  • \$\begingroup\$ The RCD/RCCB is a good idea, and I had never really considered it. In hindsight, it should probably be the first thing on an electronics bench. \$\endgroup\$ – Indraneel Apr 12 '18 at 3:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.