Multiple regulated voltages

I need +/-10V (about 1A), +/-2.5V (about 100mA) and 5V (about 60mA). I was shocked that it is virtually impossible to find a negative voltage LDO capable of hundreds of milliamps and reasonably cheap. It seems that producing all these rails with a bunch of opamp-boosted emitter followers will be cheaper than a single adjustable negative LDO. I already have a decent 2.5V reference (LTC6655) in the project so why exactly should I avoid the temptation of making this PSU out of discrete elements? Noise properties should be comparable to that of an LDO and I can shape thermal/max voltage issues with a proper transistor.

Best regards, Piotr

Reply to
Piotr Wyderski
Loading thread data ...

I'm surprised you want to waste 1.8 watts on 5V and 2.5V linear regulators. It's easy to make negative buck converters from positive boost converters.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

The fans consume 65W, so there is no real reason to care about 2 additional watts. And the rails will be quiet, which is a nice feature in the vincinity of a 24-bit ADC.

Best regards, Piotr

Reply to
Piotr Wyderski

What do you have for input voltage/current? I guess the LM337 doesn't count as LDO. Two wall warts for the +/- 10V? There are also a number of DC-DC converter modules. But 1 amp may cost too much.

George H.

Reply to
George Herold

Winfield Hill wrote in news: snipped-for-privacy@drn.newsguy.com:

Our oem lvsmps with multiple outputs were practically the same as the single output jobs. Usually a bit smaller than a brick.

We made about 400V and then downconvert for the actual output voltages. Running the 400V rail clean made the supply LV setpoints easier to manage and keep clean.

Reply to
DecadentLinuxUserNumeroUno

Why would one need an LDO to go from -10V to -2.5V?

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Some people have taken to calling any linear regulator an LDO.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

Grin you wouldn't. I assumed the LDO was for the -10 V.

George h.

Reply to
George Herold

On a sunny day (Thu, 13 Jun 2019 20:14:06 +0200) it happened Piotr Wyderski wrote in :

As long as you do not specify tolerances and ripple, you may as well use a tapped transformer a few diodes and capacitors and some oscillator. That is how PC power supplies work.

Reply to
Jan Panteltje

Right, I have no idea what an LDO is or isn't. I use the LT3080, called an LDO, but typical app has a drop out of 1.2 V.

George h.

Reply to
George Herold

it is a bit special, it is emitter follower output so in principal not an LDO, but it has extra input for a higher control voltage so it can have a low drop out

Reply to
Lasse Langwadt Christensen

Exactly. And it turns out there are no such fixed voltage LDOs, curious on its own. There are many for +10V.

Best regards, Piotr

Reply to
Piotr Wyderski

These rails are needed for a 16-bit DAC and 16 24-bit sigma-delta ADCs (2x ADS131E08). PSRR is decent and I don't insist on the last possible resolution bit, but nonetheless it would be best to keep the PSU quiet.

One strange discovery is that I can buy all the components for such a 5-way PSU at the price of a negative LDO for the -10V rail.

Best regards, Piotr

Reply to
Piotr Wyderski

LT3439, located far, far away and shielded by a several centimeters thick aluminum block (227mm, but with slots).

Not even close to the required insulation.

Best regards, Piotr

Reply to
Piotr Wyderski

To be honest, heatsinking is not exactly a problem in this case:

formatting link
formatting link
formatting link

5kg of high-performance alu + some 800 cubic meters per hour of forced airflow. I am enchanted by the sound of this unit at its full rated speed. Now just don't want to mess the analog front-end.

Best regards, Piotr

Reply to
Piotr Wyderski

What in the world are those fans cooling?

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Winfield Hill wrote in news: snipped-for-privacy@drn.newsguy.com:

Probably better off putting two of the CPU self contained water coolers under a quarter inch thick Al plate with the parts on it.. Quiet, small fans, and extremely good cooling characteristics. Much smaller end product. Oh, and the fans and radiator are remote to the actual cooling plates. Far easier to enclose and service.

Quieter electrically too. Especially proximal to the electronics.

Reply to
DecadentLinuxUserNumeroUno

LM1117 has 1.2V dropout in the worst case and it is counted as an LDO. One VBE drop is absolutely fine.

Best regards, Piotr

Reply to
Piotr Wyderski

A mother of Godzilla dual linear electronic load.

Best regards, Piotr

Reply to
Piotr Wyderski

That was one of the considered options, but a more conservative design won. You have to dump the kW-level amount of heat somewhere anyway and it is better not to have any fluids nearby when it is operating in HV mode.

It turned out it is not exactly an advantage. It is very quiet at moderate loads (the fans are PWM-capable) and you can hear every watt of its capabilities at full load. This audio feedback is in fact a post-design discovery that turned out to be a very useful and desired feature. I would do it consciously that way in a fresh design.

And sexy. More or less this pitch:

formatting link

Best regards, Piotr

Reply to
Piotr Wyderski

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.