Multiple regulated voltages

I think that's a shame, and try to avoid applying the term for anything that wants more than 300 or 400mV. We call them "lower dropout" in our book. Also see Figure 9.11, where there's a region well below LM317 types, that should be reserved for the LDO label.

Diode rectification loses about half a volt of diode drop.

Active rectification with MOSFETs can drop a lot less voltage, and you can actively regulate the rail you need to be precise, and the rest won't be too far away from nominal.

Ripple is a probelm, but post-rectification filtering and carfeful board layout can keep it pretty low.

A heat pipe might be even better. The water gets circulated as steam - so you don't need a pump. It's got to be a made-for-the-job design, but getting one designed isn't a big or expensive deal.

On a sunny day (14 Jun 2019 04:10:14 -0700) it happened Winfield Hill wrote in :

Agreed.

Agreed or not, that's what the TI datasheet says. I am not influential enough to make them consider the change seriously.

Best regards, Piotr

On a sunny day (Fri, 14 Jun 2019 15:29:08 +0200) it happened Piotr Wyderski wrote in :

Oh I would not even try many people and organisations say many things Now Iran has hit some ships or so says US Precedent. A previous one saw weapons of mass destruction in some fish and chips stands.

I do go by what I measure.

It is simple to do a discrete one with the pass transistor the other way around, so you get at most a few hundred mV drop, driven from a diff amp.

Can it do AC?

On a sunny day (Fri, 14 Jun 2019 10:37:15 -0700) it happened John Larkin wrote in :

It is probably one of those super powerful audio amps for in car stereo :-) Sometimes they 'driveby' here...

John Larkin:

Yes, that's the plan. Dual independent positive DC channels with common GND, a single combined DC with double the current capacity or AC. Plus a lot of internal and external triggers for the purpose of transient response checking and allowing the device to act as a curve tracer for magnetic components. All that with just 2 BNC connectors, PSoC5LP allows obscenic level of dynamic analog reconfiguration.

This configuration allows for a significant analog resource sharing, so there is no dual fully independent mode, with split GNDs. I don't work with bipolar sources, so I have decided to implement this simpler variant.

Best regards, Piotr

I think that the VBE should be a fair discrimination level. If below that, you are an LDO, if above or equal, sorry, no banana.

Best regards, Piotr

But seriously, what can be the rightful application of such a crazy heatsink? My seller didn't know, it was a bargain purchase for about 17 USD.

Best regards, Piotr

On a sunny day (Mon, 17 Jun 2019 09:27:35 +0200) it happened Piotr Wyderski wrote in :

Yes, I bought his one some years ago:

On a sunny day (Mon, 17 Jun 2019 09:24:10 +0200) it happened Piotr Wyderski wrote in :

Agreed. Vcesat or MOSFET will allow LDO. Anyways the ripple must then be guaranteed not to go below that. As electrolytic caps age and have large tolerances, ... a huge margin is always needed.

In the sixties I did a regulator that measured the voltage across the serial transistor and adjusted a pre-regulator (thyristor) to always have sufficient voltage to filter the ripple.

Do people still use that electrochemical fossil? (kidding)

Best regards, Piotr

We make some resistor simulators, but signal level, not power. We simulate RTDs and thermistors and such. Bipolar and wideband, since some people excite sensors with AC, or multiplex the excitation and sensing stuff.

Resistor simulation turns out to be one of the most difficult things we have done, especially because we need isolation and serious accuracy.

P620 here

and a VME module

We have a couple of old Leader active DC load boxes, which are really great to have around once in a while.