Linear transformer repetitive switching.

[davelectronic]

Recently I've had an idea for a power supply, but rather than being on, it would be on standby at a ready state. I was thinking of trying to use a linear transformer power supply, but it's not on until a button is pressed, and then it's ready to power the load. The load would be a HF amplifier. The idea behind the PSU not being on until demand for power is switched is to save power and wasted heat. A few problems I'm unsure about is repetitive switching ( several minutes between on off periods) and inrush current, and possibly some latency. As it's fast to power on, but not that fast. I would use a lower standby transformer to keep the amplifier on all the time, but that's only a few hundred milliamps if that. My plan was for a relay (high current) to switch on the high power linear transformer. Do you think this a feasible idea, or is it likely to be fraut with problems as mentioned above. Any thoughts and ideas appreciated, and thanks for reading.

[WattsThat]

Makes no sense as explained. The only difference is the magnetizing currents in the transformers.

[coromonadalix]

The only thing who would be good  is to have switched secondaries  to have like  half of the normal supply voltage going into it ....  you have many bench psu who use this principle,  mine too

If not,  use an smps  variable psu,  you'll cut your ''losses'' big time

But  if you only want to reduce the power up times and save lost energey / heat  etc ...  it's kinda too much ??

For the power supply in question, what is it ...

[T3sl4co1l]

So, keying the finals by keying their power supply?  (Or in addition to.)

If the amp isn't consuming any power when keyed off anyway, it doesn't make a difference.  (Presumably you already know it doesn't.)  I suppose a poorly designed might still pull bias current?  Or a more complex one might still power preamp/mixer/exciter/driver/etc..  Or a tube amp will still be pulling heater, which I think you'll prefer (simmering) for instant-on..?

A switching supply, you can turn on and off nice and quickly (within ~ms), but a linear supply will take multiple cycles for its filter to discharge, long enough you may find it unappealing to cycle.  It also draws a big gulp of inrush when turned back on again, and if it was operating under load quite recently, NTC inrush protection won't be ready for that (if so equipped).

Tim

[bdunham7]

Recently I've had an idea for a power supply, but rather than being on, it would be on standby at a ready state. I was thinking of trying to use a linear transformer power supply, but it's not on until a button is pressed, and then it's ready to power the load. The load would be a HF amplifier. The idea behind the PSU not being on until demand for power is switched is to save power and wasted heat. A few problems I'm unsure about is repetitive switching ( several minutes between on off periods) and inrush current, and possibly some latency. As it's fast to power on, but not that fast. I would use a lower standby transformer to keep the amplifier on all the time, but that's only a few hundred milliamps if that. My plan was for a relay (high current) to switch on the high power linear transformer. Do you think this a feasible idea, or is it likely to be fraut with problems as mentioned above. Any thoughts and ideas appreciated, and thanks for reading.

So I assume this is still the project with MOTs?  I think if you do this right, it will work fine.  You just have to make sure that you have a fast enough relay (single-digit ms) and large enough filter caps to power the load for at least a full cycle, preferably a bit longer.  The transformer isn't going to care much about being switched on and off and the surge currents should be small if the filter caps are already mostly charged up.

I have some high-quality audio amplifiers that do something very much like this.  They have a 'boost' mode that is engaged by a relay if the power demand goes over a nominal level.  It just changes taps on the transformer. It functions quickly and without any noticeable disruption in the operation.

Reading other replies, it isn't clear that we are all assuming the same thing here.  I'm assuming you have a load (amplifier) and you want to keep it powered at all times with essentially two power supplies in parallel, the smaller one being on all the time and the larger one kicking in when you 'press the button'.  You would use a small and a large transformer, each with its own rectifier bridge, connected in parallel to the filter caps of the power supply.  If that's not right, my answer is wrong.

[Benta]

The big/small transformer doesn't make sense either.
Actually the opposite.
Magnetically a big transformer is much more efficient magnetically than a small one (lower loss).

The point is, that the primary magnetizing current is "free of charge", except for cable losses, but they are minimal. So are copper and iron losses at this low primary current.

Keep the primary on and switch the secondary(ies) according to load needs.

[bdunham7]

Magnetically a big transformer is much more efficient magnetically than a small one (lower loss).

Ordinarily, yes.  But IIRC, his power supply project is using reworked microwave oven transformers. 

[Benta]

Magnetically a big transformer is much more efficient magnetically than a small one (lower loss).

Ordinarily, yes.  But IIRC, his power supply project is using reworked microwave oven transformers.

OK, didn't know that.

[T3sl4co1l]

The big/small transformer doesn't make sense either.
Actually the opposite.
Magnetically a big transformer is much more efficient magnetically than a small one (lower loss).

The point is, that the primary magnetizing current is "free of charge", except for cable losses, but they are minimal. So are copper and iron losses at this low primary current.

Keep the primary on and switch the secondary(ies) according to load needs.

Well, the absolute losses will be more -- there's more iron to magnetize. Losses can be lower with suitable design, but that would waste space/cost, so you're not going to see it off the shelf anyway. (I mean, unless you go out of your way to use a 240V unit at 120, with ~4x excess capacity.) But relative losses are generally better, yeah. If not in this particular case with an MOT, heh.

Tim

[Benta]

Well, the absolute losses will be more -- there's more iron to magnetize. Losses can be lower with suitable design, but that would waste space/cost, so you're not going to see it off the shelf anyway. (I mean, unless you go out of your way to use a 240V unit at 120, with ~4x excess capacity.) But relative losses are generally better, yeah. If not in this particular case with an MOT, heh.

Tim

I agree that most transformer manufacturers these days drive the cores to the limit, but there are exceptions. That costs more, of course.

But with MOTs, all bets are off.

[coromonadalix]

A toroidal xformer will be the best,  more efficency and less core losses

[davelectronic]

Thank you for all your replys, very interesting. Bduham7 has it right or closest to what I'm trying to achieve. The power supply with a single MOT (microwave transformer) gets stupid hot when just idle, I've overcome this with two MOT's in series, each sharing the 240 Volts AC mains input. This does drop the secondary voltage AC by half, so I needed more windings secondary to get the secondary AC voltage I need, about 18 Volts AC. The idea of having a single MOT power ready but not active whilst there's no demand for power. Switching in the MOT to power the amplifier, but then when not powering the amplifier the standby transformer keeps the amplifier powered on ready for the heavy power drawn from the MOT for the amplifier. Hopefully that makes sense, a single MOT of 700 watts draws 3 Amps when it's just idle powered on. The two MOT in series draws 0.185 Amps when there idle. I'm trying to find a solution for a single MOT, and have a solution to the overheating. Any alterations to a single MOT primary just won't work due to limited space for extra windings on its primary side. I've been using a two MOT in series for over 6 months continuous now, it's 12.60 Volts DC at up to 30 Amps. And it works very well. Now I'm trying to find a solution to using a single MOT transformer to achieve the same thing, but with out continuous wasted heat and power. The amplifier is roughly used on a 50% duty cycle for RX to TX states. My idea might not be a feasible one, but there might be a way to achieve this.

[NiHaoMike]

I've been using a two MOT in series for over 6 months continuous now, it's 12.60 Volts DC at up to 30 Amps.

Better to modify a PC power supply for that.

[T3sl4co1l]

Try 2.2uF in parallel with the primary?

(It's only 44 VA, and it's not obvious whether it's due to magnetizing inductance or actual core losses.)

Tim

[davelectronic]

The 3 Amps current of the 700 watt MOT I took by putting a meter in series with the transformer. I was quite surprised as there was not secondary load on the transformer. I did load it up with about 300 watts of halogen lamps, the secondary AC didn't drop hardly anything. Seemed a very solid secondary output. Heating with no load reached 90° C I wasn't going to let it climb any higher. I know the insulation class is 200° C but that's dangerously hot. 2 X MOT's in series with a 20 Amp load continuous gets to about 70°C which I thought exceptable. Also being as I don't need continuous operation at full load. Is the capacitor something akin to a motor start run capacitor ? Altering the power factor. I'm not sure if that's the idea, but I will try it, I need to get a capacitor of that value as I don't have one. The pc power supply works just fine for this application of powering a linear amplifier, I've just got the BIG linear power supply mission in my head.

[T3sl4co1l]

Yes, MOTs get rather hot by themselves; that's the point, intermittent use.

How hot did they get with the two series configuration, no load?

70C is a, whatever, 45C temp rise, quite acceptable for full load operation, yeah.

The capacitor will null reactive current, on the assumption that the unloaded current draw is entirely reactive power.  Whatever's left will be resistive loss, which you can then measure as current draw.  P = S when PF = 1, and S = |V| * |I|.  (You need a power meter to measure \$P = \overline{V I}\$.)  Ideally, you'd use a range of capacitors, say from 1.5 to 3uF, adjusting until you find the least current.  This will still measure in excess due to harmonic currents (which should be small at the low magnetization, but won't be quite zero), so it's still not exact, but it should be much closer than taking raw primary current times mains voltage.

Tim

[bdunham7]

The capacitor will null reactive current, on the assumption that the unloaded current draw is entirely reactive power.  Whatever's left will be resistive loss, which you can then measure as current draw.  P = S when PF = 1, and S = |V| * |I|.  (You need a power meter to measure \$P = \overline{V I}\$.)  Ideally, you'd use a range of capacitors, say from 1.5 to 3uF, adjusting until you find the least current.  This will still measure in excess due to harmonic currents (which should be small at the low magnetization, but won't be quite zero), so it's still not exact, but it should be much closer than taking raw primary current times mains voltage.

I've always assumed that those non-interleaved butt-welded cores would have horrendous losses even if they weren't madly overdriven into saturation, but I've never checked.  I did read something recently that claimed that the eddy current due to the welding was no big deal, but I'm not sure why.  It would be interesting to see exactly what the V/I relationship is on his half-voltage-primary MOTs as well as how low he can get the recirculating current. 

[T3sl4co1l]

It's not as bad as it might look; it's around the edges only, so doesn't make a full shorted turn, just a comb shape.  There are some eddy currents in them directly (cuz yeah, there's field going right up by/into the weld zone), but the widest loops have to trace a circuitous path around the outside, back around another weld, or something like that -- far higher than a direct shorted turn kind of loop would be.

Tim

[davelectronic]

The two MOT's in series 240 Volt AC input, each transformer receives half the mains voltage. This impacts the secondary windings, I needed at least 18 Volts from the secondary windings. Each secondary gives 9 Volts AC or just over that. 14 AWG is as heavy as I could go, only because of area for winding the cable on. The no load temperature is around 48°C fluctuating with what the rooms ambient temperature is. I think I understand the AC capacitor theory, and I'm keen to try it, two MOT's in series for a high current low voltage works very well, my current radio set up has been using this series MOT power supply for over 6 months. It on 24/7 no problems at all, just the added weight of that second transformer so they function with out overheating. The unit has fans, but there only on when the PSU is working heavily. The cable I rewound the transformers with was silicone insulation sheathing. I used a blunt round nose follower, like a marlin spike but rounded polished end to get more turns on, silicone spray made it relatively easy to get the extra turns on. A few years ago I did try splitting the IE core, but once reassembled the no load current was worse than 3 Amps idle, some how it must of upset the magnetic circuit. I did think the welds equalled a shorted turn, I did realise the shallow depth loop it created, I know these things aren't perfect, but the price of high VA transformers is rediculous. And the MOT transformers to me where free, I only had to buy the secondary cable. The unit in use is in the picture below, the two transformers and bridge rectifier are in the lower half of that power supply. Everything else is in the top half of these two cases bolted together. It's very heavy, I did weight it, 10 to 12 KG from memory. That's why I'm chasing a single MOT variant.