Adj. +/-60V to +/-70V DC Power Supply

[onemilimeter]

I wish to build a +/-60V ~ +/-70V DC power supply which is capable of supplying 10A. What are the options I have?

Thanks.

[Psi]

Given that +/- 70V is a potential difference of 140V and  140V x 10A = 1400W that's a pretty powerful power supply.

It will have to be switchmode, atleast for the most part, you might be able to do a hybrid system and have small amount of linear regulation on top of the switchmode output if you absolutely need low noise.


If i was doing this i'd probably use a large toroid transformer from a 1-2kW audio amp. They tend to be 230v/110v in and +-70V output, or around that.
Then i'd do a inductor based switchmode on each 70V side to create the 0-70V voltage adjust.
The disadvantage of doing it that way is the size and weight of the toroid, since they're mains frequency they huge and heavy for that kind of power.

The other way is to do a full switchmode system and use high frequency for the whole thing, that way you can make it smaller and lighter. However it might end up being quite tricky to design a 10A 1.4kW power supply at 230V input. Its also not as safe as you working with 230V instead of the 70v out of the mains toroid

[onemilimeter]

Given that +/- 70V is a potential difference of 140V and  140V x 10A = 1400W that's a pretty powerful power supply.

It will have to be switchmode, atleast for the most part, you might be able to do a hybrid system and have small amount of linear regulation on top of the switchmode output if you absolutely need low noise.


If i was doing this i'd probably use a large toroid transformer from a 1-2kW audio amp. They tend to be 230v/110v in and +-70V output, or around that.
Then i'd do a inductor based switchmode on each 70V side to create the 0-70V voltage adjust.
The disadvantage of doing it that way is the size and weight of the toroid, since they're mains frequency they huge and heavy for that kind of power.

The other way is to do a full switchmode system and use high frequency for the whole thing, that way you can make it smaller and lighter. However it might end up being quite tricky to design a 10A 1.4kW power supply at 230V input. Its also not as safe as you working with 230V instead of the 70v out of the mains toroid

Thanks. As the device is intended for laboratory test, weight or efficiency is not a prime concern at the moment. Thus, I prefer the first approach you suggested.

[johnwa]

How well does it need to be regulated? How about putting a variac in the transformer primary? I have a smaller unregulated supply (20V 10A) that uses this arrangement, it is pretty indestructible, and is good for things like high power motors. You could add a tracking linear regulator, or possibly look at feedback controlled motorised adjustment of the variac if transient response is not a concern. Sorry if this isn't a very 'electronic' solution, but depending on your requirements it might turn out to be easiest.

[Zero999]

55V-0-55V transformers are widely available which will give you an unregulated power supply when combined with a variac, rectifer and capacitor.

You could build a linear regulator, put it on the output and mechanically couple the adjustment potentiometer to the variac so the input voltage tracks the output voltage, which will give you reasonable regulation and efficiency.

[alm]

If the spec is for a symmetric +/- 60V DC to +/- 70V DC, linear actually wouldn't be too bad, since you only have to dissipate 2 * ~10V (+ ripple) * 10A, which is not too bad. For 0 to +/- 70V DC, I would either go switching or use some sort of pre-regulator (variac, tap switching, SMPS or triac). It also depends on the requirements for regulation, noise and overload protection.

[onemilimeter]

How well does it need to be regulated? How about putting a variac in the transformer primary? I have a smaller unregulated supply (20V 10A) that uses this arrangement, it is pretty indestructible, and is good for things like high power motors. You could add a tracking linear regulator, or possibly look at feedback controlled motorised adjustment of the variac if transient response is not a concern. Sorry if this isn't a very 'electronic' solution, but depending on your requirements it might turn out to be easiest.

I need them to supply a linear audio amplifier. Thus, I hope the "dc" voltage is as stable as possible...

[onemilimeter]

55V-0-55V transformers are widely available which will give you an unregulated power supply when combined with a variac, rectifer and capacitor.

You could build a linear regulator, put it on the output and mechanically couple the adjustment potentiometer to the variac so the input voltage tracks the output voltage, which will give you reasonable regulation and efficiency.

I'm aware of 7800 and 7900 series linear regulators. Do we have high output voltage linear regulator in the market? You suggested that "you could build a linear regulator". Did you mean to build one from discrete components? Kindly share how to construct one, e.g. schematic. Thanks.

[alm]

I need them to supply a linear audio amplifier. Thus, I hope the "dc" voltage is as stable as possible...

Non-switching power supplies for power amplifiers are typically not regulated, just a transformer, bridge rectifier and lots of smoothing capacitance. The power supply ripple rejection of the amplifier does the rest. The exact output voltage doesn't really matter in that case, so it's fine if it changes depending on the mains voltage.

I'm aware of 7800 and 7900 series linear regulators. Do you have high output voltage linear regulator in the market? You suggested that "you could build a linear regulator". Did you mean to one from discrete components? Kindly share how to construct one, e.g. schematic. Thanks.

There is the LM317HV, but I don't think it goes to 70V. If you don't mind the regulator going up in smoke if you overload it, you can raise the potential on the ground/adjust terminal so the regulator sees less voltage. It doesn't really matter, since I doubt you'll find a monolithic linear regulator that can supply 10A at 60-70V, so you'll need a bunch of discrete components anyway (although you may be able to use a monolithic regulator with external pass transistors, if you don't require current limiting).

[Zero999]

Yes, there's no point in using a requlated power supply for an audio amplifier. A linear regulator will just waste power and an SMPS could produce noise which might interfere with the amplitude. Just use a centre tapped 100V or 110V transformer, a bridge rectifier and some large capacitors. You probably also need a soft start but that's pretty easy to make using a relay and comparator or 555 timer.

[onemilimeter]

Thanks alm and Hero999...

[Psi]

here's an example of the toroid in a high power audio amp. No idea of the exact power of that one but it looks like 1-2kW

To give you some idea of the size, thats a 19" rack mount box, probably 4U high

[onemilimeter]

Found a power supply circuit (see below) in the datasheet of a linear power amplifier. Why does the circuit need two 500ohm resistors at the output stage? Thanks.

[andyg]

Found a power supply circuit (see below) in the datasheet of a linear power amplifier. Why does the circuit need two 500ohm resistors at the output stage? Thanks.

It's there to bleed/discharge the filter capacitors after the unit is turned off. Otherwise the caps could still be charged when you turn the unit off and you'll probably get unintentionally zapped one day.

[Zero999]

It doesn't need any resistors, they just waste power. The only reason why one would connect resistors across the supply is to discharge the capacitors but 500R is far too lower value, they'll waste 20W when the power supply is running. If the audio amplifier is connected to the output of the circuit, the capacitors should discharge to a safe voltage in a couple of minutes.

EDIT:
You need a soft start, otherwise it'll blow the circuit breaker when you turn it on. Here's an idea which also discharges the capacitors when turned off. When initially powered on R1 limits the current to the PSU, after a second the 555 timer turns K1 on short circuiting R1. When powered down R2 is connected across the output to discharge the capacitors. The circuit could be improved by using a separate relay for R2.

[onemilimeter]

Thanks andyg and Hero999 for clearing my doubt and for the "soft-start" suggestion...

[onemilimeter]

Hi Hero999...

You mentioned about "soft-start" for the transformer circuit in your post (https://www.eevblog.com/forum/index.php?topic=2457.msg33609#msg33609). Without the "soft-start" you said it would blow the circuit breaker when it's turned on. Is the "inrush current" due to the "magnetizing branch (Lm)" of the transformer or due to the capacitors at the output stage or due to the load if a load is connected? Will we get the same "inrush current" if the secondary of the T1 is left open?

Thanks.

[Zero999]

Both the inrush current from the transformer and the capacitors can cause the fuse to blow.

[NiHaoMike]

The most usual solution for limiting inrush current is a series NTC. If needed, bypass it with a relay to eliminate the slight efficiency loss at high output current, but beware that at low current, the relay would likely use more power than the amount saved by bypassing the NTC.

And at high powers (more than 25W or so), switching power supplies are a better choice. For low noise, add some LC filters and maybe a LDO postregulator.