managing huge trafos

[nania]

...so in this case soft starting the primary can be mandatory...

...and I can be a moonbeam out on a star...
What's the final verdict, will I need inrush treatment on the primaries or not if the transformers are isolated? Thanks for pointing me to magnetic starters, they seem appropriate here. Once I get the supply juiced up, what's the best strategy for keeping it there in standby?

This relay does not switch off the amplifier and thus does not see the worst of the wear/deterioration that the ON/OFF switch/relay sees.

Yes, but it's still in the circuit and it can still see 30A. Ideally, I'll need a relay that matches the settled thermistor resistance and those are usually beefier than milliohm types. Speaking of settled thermistor states, 2.5R is not available and ST1R30B seems to be the only one that can handle 30A and it's only available in 1R according to this vendor. I can put them in series to get a higher value with the commensurate power penalty but is there a more suitable part for this application?

 

[AndrewT]

Yes, you can put two or more NTC power Thermistors in series.
The resistances add up. The joules energy absorption also add up if they are the same value.

Do not put NTCs in parallel.

 

[TonyTecson]

i will use a 10 ohm 100 watt Dale metal clad power resistor shorted out by a relay contact after a second delay..

i never used NTC's if those are to be shorted out, i keep them in circuit all the time....and i have seen failures of NTC's in amps that used them, that is why sales of NTC's here are brisk....
NTC's when used in series needs 1meg equalising resistors, and when used in parallel can suffer current hogging, aster all they are semiconductors....such a low mass makes them susceptible to failures...

bottomline, i never used NTC's if i have to use them in series or in paralel....for it is just one NTC or none at all, my 2 cents...

An externally hosted image should be here but it was not working when we last tested it.

youtube tutorial....Step-Start (Soft-Start) Relay Circuit For High Voltage Power Supply & RF Amplifier - YouTube

 

[trobbins]

Isn't the simplest technique to switch the transformer on 10 times and see if any cb or fuse trips, or any other annoyance occurs???

If a cb trips then check its rating and trip curve - for example you may need to swap out a C-curve for D-curve.

Why go to all the other bother unless you have an actual valid reason.

 

[martin clark]

Mmmm; maybe; perhaps the new toy will survive, but other things on the same phase, not so.

Which brings up another point:


Snubbing.

With such an energetic, inductive load (large potential I , therefore very large 'potential' from L.dI/dT at turn-off), DO add a snubber across the input windings next to the transformer, i.e after the switching mechanism: so that when you de-energise, you also get a graceful shut-off without arcs/sparks or at least EMI issues elsewhere.

In the UK LCR produce nice integrated Class X -rated parts that are 100nF+100R in series, or 220nf+100R in series, in a single pack; cheap, too:

Encapsulated RC Suppressor - LCR Capacitors

(this or similar strongly recommended; in UK, these are available to DIYers from Rapid Electronics, cheap, and even from Maplin - part RG22Y)

 

[trobbins]

I'd recommend installing a MOV across the primary winding as a simpler form of snubber - for the reason's Martin suggests, as it is a way to alleviate switch arcing, and over-voltage stress on the primary winding and all related wiring/parts.

There certainly could be annoyance to other equipment by the switching action of the power transformer - such as audible clicks in music gear, or a transient loss of TV image, or a flicker from lighting - it all depends on the site and equipment, and may already be a concern when other equipment turns on/off, such as a refrigerator or freezer or a direct on line motor.

 

[TonyTecson]

i use MOV's after the fuses as a matter of practice....a good practice i would say...

 

[nania]

TonyTecson
I'm planning on fusing the secondaries to make them part of the pi filter. Is there any advantage to fusing prior to the juice entering the trafo? How do you completely short out the resistor without adding diodes? If you parallel a second conduit, isn't the resistor still in the circuit burning watts? Thanks for the link, is the circuit in that video the one you use?

trobbins
Yes, the MOV is a practical inclusion, are there any that damp the energizing inrush with a 5kW trafo so that the soft start/standby circuit can be placed after the secondaries?

martin clark
Thanks for the link.

 

[TonyTecson]

TonyTecson
I'm planning on fusing the secondaries to make them part of the pi filter. Is there any advantage to fusing prior to the juice entering the trafo? How do you completely short out the resistor without adding diodes? If you parallel a second conduit, isn't the resistor still in the circuit burning watts? Thanks for the link, is the circuit in that video the one you use?

yes, go ahead and fuse the secondaries, and primary too....
or you can use resettable overcurrent current fuses on the primary side..
fuses in the secondary side must not interfere with the normal working of psu, or they are not meant to prevent over current in that sense, just to prevent catastrophic failures, as you gain more confidence in your build you may find that they are not absolutely needed..

An externally hosted image should be here but it was not working when we last tested it.

i posted the video to give you the idea on how soft starting is done and basically it....
you need dc to energize the relay hence the diode, the series resistor after the diode and the capacitor caused the delay, after that the power resistor ib series with the traffic primary is shorted out...

 

[stocktrader200]

if you use a zero crossing circuit with relay, that should limit the trafo current due to core energizing. if you are charging a large bank of capacitors on the secondary, you will need a 150 - 200w resistor and a second relay to bypass this part after several seconds.

 

[TonyTecson]

if you use a zero crossing circuit with relay, that should limit the trafo current due to core energizing. if you are charging a large bank of capacitors on the secondary, you will need a 150 - 200w resistor and a second relay to bypass this part after several seconds.

i was chastised for saying that in another thread many years ago...
turn on current surge it seems is maximum at the zero voltage crossing, not good...

 

[trobbins]

Yes, the MOV is a practical inclusion, are there any that damp the energizing inrush with a 5kW trafo so that the soft start/standby circuit can be placed after the secondaries

You should aim to appreciate how a MOV works before installing such a part. They clip over-voltage peaks - which may arise from switching current through a winding due to its inductance. A MOV won't 'damp' inrush.

I'm planning a universal bench supply so amps are only one of the things it will see.

I'd recommend you appreciate how to over-current protect both the primary and secondary side of any transformer - that is not a trivial process, and somewhat depends on your application (which may not be well defined for a general purpose supply). As you are getting up in power level, that may mean other protective parts than just fusing may be as practical, or more available to you. One example article on secondary side fusing for amps is:
https://www.dalmura.com.au/static/Valve%20amp%20fusing.pdf

 

[martin clark]

Correct, Tony: switching transformers at the zero-crossing turns-out to be a rather bad idea.

Example:

http://www.te.com/commerce/Document...v&DocNm=13C3206_AppNote&DocType=CS&DocLang=EN

(- I have other papers too)

 

[TonyTecson]

A MOV won't 'damp' inrush.

yes, and their small mass can only take so much and they too will fail if not used judiciously,

just like NTC's, i once had a pc that makes a loud bang on power on and yet continues to load windows and appear normal, after several such episodes it gave up totally...a post mortem revealed a broken in two pieces along a deep line in the middle of the disk...

to sum it up, MOV's are shunt devices and NTC's are series devices wrt power traffos...

 

[TonyTecson]

Correct, Tony: switching transformers at the zero-crossing turns-out to be a rather bad idea.

Example:

http://www.te.com/commerce/Document...v&DocNm=13C3206_AppNote&DocType=CS&DocLang=EN

(- I have other papers too)

thank you, much appreciated...please link the other papers if related to this this topic....

 

[martin clark]

That one is the most concise.

This is also a good read:

http://emeko.eu/fileadmin/_migrated/content_uploads/02-Press-TSR-Speech-cwieme2004-e.PDF

HTH.

 

[TonyTecson]

thanks martin, indeed careful planning and implementation of power supplies can yield positive results....the best psu so far that i did was for a 6C33C set, the psu yielded a power factor of 100% measured at the primary side when the amp is all powered up and playing music....

i could not replicate that since...

 

[nania]

Thanks for the tips guys. I was hoping to have just one inrush limiter but the opinions indicate I'll need one on both the trafo input and output. Why doesn't the inrush limiter on the trafo output also slow it's input inrush?

While that's mulled over, it looks like a 100A rated solid state relay with a resistive heater coil is the simplest and most economical solution to the energizing inrush. These relays come in zero and random crossing type. Does the zero crossing type solve the issues described in the papers linked by martin clark above? I suppose they could by 'forcing' or waiting for the zero crossing state but I'm not sure. I am planning on mounting the relay on one of the aluminum case sides which should be able to dissipate 5-10W safely. I expect a 2mA leakage from the relay, am I safe with a 30A demand?

 

[TonyTecson]

a zero crossing ss switch has less emi/rfi generations from what i gather...
a random switching works just fine, MOV and snubbers shunting the transformer primary helps too..

 

[trobbins]

I was hoping to have just one inrush limiter but the opinions indicate I'll need one on both the trafo input and output. Why doesn't the inrush limiter on the trafo output also slow it's input inrush?

The transient peak due to magnetising current of PT will supress the starting portion of charging up any secondary side filter capacitance (when using ss diodes) and valve heater loads, as power transfer through the transformer is restricted. But once magnetising current starts to subside, then the charging of secondary side caps still exists, so the mains AC supply continues to see high current peaks.

Any over-current protective device on the ac mains will trip if their I2t current-time curve is exceeded. Fuses and CBs can be selected to allow larger/longer current during such turn-on events, so you may not need special limiting schemes at all.