• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Thermistor protection during brownouts

I plug my power strips or the individual pieces of equipment into them. I have 5 total. They work great at protecting my equipment from hot-restarts which is what I want. Not overly sensitive in my situation. Never have they tripped for a fault in my equipment. Well, the one on my bench main power strip has plenty of times during new builds or experimentation. How do you know that mine trip at 4-5mA imbalance?
Just like THIS
When looking up GFCIs I noticed that only the current rating of 15 A was given, but never a current imbalance (ground fault) tripping rating. So I assumed that they don't need the rating specified if they are all the same. I remembered the 5 mA, probably from a YT video (possibly Electroboom). A search confirmed they typically trip at 4-5 mA of imbalance.
Our RCDs always have the ratings specified (printed on them) as you can have different types, and values of imbalance tripping and contact ratings. The 5 mA types are sometimes used in schools or labs to protect sockets on the work benches.
Over time I found the thermistor in-rush limiter isn´t reliable. I always use a high power 25-50W 25-50 Ohm resistor in series with the mains transforrmer (toroid or high VA E&I), timed relay switched out after the surge has passed. I´ve saved countless fuses over this method. It isn´t perfect but is an improvement. The cue to success is a fast reset circuit and proper fuse rating which avoids the worst of that damaging back-flux which basreflex describes. That is especially noticable with intermittant loose (worst case) or poorly wired utilities, where the relay reset must occur on the instant of AC waveform interruption. As another has described, a triac switched out is also possible but more circuitry is unavoidable.

Ingenuity holds no bounds.
Bench Baron
From what I found, a brown out is an intentional voltage reduction by the utility company
Not necessarily, all depends on where you live. I used to live in an old neighborhood with an SWER power system and during the summer the voltage was always far below spec, and the power company wouldn't, or couldn't do anything about it. SWER stands for Single-Wire Earth Return. One hot lead fed the pole transformers and the return to the power station was through the earth, so the voltage regulation is terrible. Sometimes florescent lights wouldn't start or the fridge compressor would stuggle to get going. I measured voltages as low as 100 VAC several times. I moved 10+ years ago but AFIK the system is still the same. I know it's rare but it's still out there even in this day and age, and especially since it was in a large metroploitan area and not the rural countryside.
But that is different from a brown out. From what I found, a brown out is an intentional voltage reduction by the utility company of 10-25% to prevent a blackout.
No, a brown out is not 'intentional', but rather it is a typical response of the distribution network due to abnormal loading somewhere, which could even be within your house, or on the other side of the country, or city or suburb or street. It is also not characterised by '10-25%', and does not 'prevent' a blackout.
The info I got on brown outs being intentional you can read here, but I see interpretations differ. Maybe the meaning is different in different countries?

Brown outs aren't a thing where I live as the grid operation is strictly regulated by laws that are enforced. Unfortunately, it also means that expansions and new connections are hardly possible anymore as our grid has reached capacity at peak times. Waiting lists are in the hundreds to thoudsands of applicants. To prevent brown outs or black outs congestion management has been set up.

In the mean time, the TSOs and DSOs are upgrading the grid, sometimes to a capacity of 500% of original, to cope with the increasing demand but also local generation of electricity. In the 1950s, large natural gas reserves were found, and as a result a lot of the energy demanding heating has been done with that instead of with electricity. It wasn't until recently that the grid operators realized that we are now lacking capacity for the transition away from fossil fuels and a lot of work needs to be done.
I wonder what the near future will bring as 'load shedding' would be the very last resort the operators over here are allowed to use.
Ta, thanks for the link. Caveat emptor - written by the owner of a small company specializing in EMC test equipment rentals, and the owner presumably has little in the way of residential/commercial/industry awareness or professional qualifications imho.

I agree that regulated grid connections, and grid management typically suppress flicker and fluctuation events, as such events are typically mandated for when approving equipment sales or installing distribution - whether residential, commercial or industrial. Most countries are playing catchup, as historic spending on poles and wires has typically been avoided. I don't recall the last residential brownout, except for flicker when a cb trips on one house feed. Rural/remote environments are very different here, as illustrated by the SWER line example and a general aging population of loads, some of which are not very nice to the mains like DOL pumps.

But I would suggest that most grievances require custom assessment, and the topic is way too broad to generalise any one 'solution'.
To make things a little more complex, the terms; under voltage, over voltage, interruption volts but not classed as a "brown out", transient spikes and so on. Important is also the fuse link characteristics, "fusible blow" times expressed as time to rupture. So ending up as T rated fuses, S slandard blow, HRC very fast, and not forgetting the earlier often used the forced protection crowbar on early TTL.

My simple basis with the relay is to avoid the punishing back emf blow that basreflex points out combines with the inrush delay function, for this I use a relay with a relatively fast drop out 5mS time coupled with the appropriate rated standard blow or Time delay fuse with an I² t rating specified on fuse manufacturer websites. I found this deals quite effectively with the loose mains connection syndrome.
There are triac circuits, but again the trigger reste voltage has to drop out even faster, otherwise the back current from an inductor could fuse the triac into a short circuit.

For everyone; I can recommend Keith Billings Switchmode power supply handbook IBSN 0 07 005330 8. Although published in the matured era of switchmode power in 1989, times have moved on, but everyone should have a copy as it defines every aspect of the line voltage issues and problems. The amount of physics in it, is easy for anyone.

Beware as always with this line voltage stuff; not for novices. Don´t get stung by unnecessary shocks and be aware of the "grasp effect" and First Aid .. Use bench isolating transformers...and importantly, keep the other hand away from completing the body circuit. I worked on MV polyphase and that will make one even more cautious.
One job I do not like is inspecting grid substation control systems when something ain´t right esp sychronization......the enormous hum and HV "sizzle"can be very intimidating.
Stay safe!

Bench Baron

Here's what EMF can do in old fluorescent lighting fictures, the type with coil and glowlight starter.
In this case the suppression caps that sit between L and N in the fixtures had become damaged by moisture. The recording doesn't do the actual bang justice, but you get the idea. Look at the end of the fixture, you'll see an arc. This arc comes from 20 cm deep inside the fixture.
Here is the weekend sift through the folders,, that inrush circuit which I cobbled up many decades ago and is standard in my mains tube/valve equipment. I even used a near identical circuit for an arc welding transformer, which is probably the worst and most hostile environment. As mentioned, all circuit values are arbitary and be changed.

Only a note, if one breadboards the circuit, which typically have poor grounds/returns, the thyristor is quite sensitive to any spikes and may latch on prematurely. I use a ground plane and keep gate components esp that capacitor typ 1-47nF close to the cathode, that way spurious operation is avoided.

Bench Baron
inrush circuit.jpg
Apols for the short component abbreviations......Naughty old habit burning midnight oil....been in this for a long time now..perhaps too long for old dogs to change habits. Zener volts merely to protect the thyristor. Can be any Zener 30-50V. At that time 1980, in non critical applications I simply used a PKE out of the junk box which is really a high speed transient type.
The great virtue of these quickie circuits is one can experiment the circuit function on the low volts side without being exposed to high volts. The mechanical relay makes an excellent interface.
stay safe

You said:
"I understand closing the power switch on a toroid power source may destroy the switch if the switch happens to be closed at the zero crossing of the line voltage."

Did you mean with the amplifier running, and then Opening the switch at zero crossing?

I say:
"Zero crossing is zero volts, closing the switch then . . . zero volts/any load resistance, no mater how low, no matter how reactive = zero amps.
And . . .
Instead, closing the switch during the max volts (at the crest, peak volts)
"At max volts, closing the switch then . . . max volts/any load resistance = lots of amps;
Especially if the secondary is loaded by a rectifier that is followed by a high capacitance capacitor input filter.

Choke input filters anybody?

For a toroid with an un-loaded secondary . . .
That takes some Inductive reasoning to analyze.
Your Mileage May Vary

Just my opinions
Last edited:
it all depends on the magnetization state of the core. switching at zero volts when the core is magnitized could add up or decrease the state of magnetization The adding could result in saturation. There is a german manufacturer of a transformer starting device. It magnetizes the core slowly with dc, then kicks in the ac at a moment such that the transformer is perfectly synchronized. no inrush current at all.
  • Like
Reactions: 1 user
basreflex and others,

With a soft start circuit:
A soft start circuit changes things. Generally . . .
It makes things Better, more complex, more parts, more chance for wiring errors, and more parts to fail.

Without a soft start circuit:
For either a toroid, double C-core, or an E-I power transformer that has a secondary which is loaded by a solid state rectifier (zero warm up time) and the rectifier is loaded by a capacitor input filter . . .
If there is low leakage inductance from the primary to the secondary, then the primary sees a Very Low impedance reflected to the primary when the power is switched on (high inrush current), during the first Alternation that goes from zero crossing to peak voltage (crest).
And, it should be even worse if the switch is closed at the peak voltage (crest).
It makes things less complex, fewer parts, less chance for wiring errors, and fewer parts to fail.
I could be wrong about this; help me to understand if it is different than I just described it.

Comments . . . Please?

Residual Magnetization:
If the toroid core was still magnetized at power down, the issue at the next switch closure would be:
The polarity of the first Alternation, which is Either trying to Increase the residual magnetization in the same direction,
Or, is trying to Decrease the residual magnetization, and then change the direction of the magnetization (more inrush current in this case).

Comments again . . . Please?

Hot Starts:
How does the soft start circuit respond to a hot start?
How does a non-soft start circuit respond to a hot start?
The response of both, may vary.

Just my opinions again
Last edited: