Maida regulator

[Brett]

I never knew these had a name. I've always thought of them as POOGE regs from similar ccts in the old AA magazines.

Anyway, excellent points SY.

[HiFiNutNut]

Thanks again, SY. I am now ready to buy the parts and build it.

The LCRCRCRC PSU has served more than a year but I have found a problem with it so it will be replaced. I am now convinced that regulated PSU is better.

Regards,
Bill


Brett: from your past posts your favourite PSU is GP's CCS - shunt reg. It would be interesting to compare the Maida with that and see if there is any difference to the 12B4 circuit.

[Brett]

Originally posted by HiFiNutNut
Brett: from your past posts your favourite PSU is GP's CCS - shunt reg. It would be interesting to compare the Maida with that and see if there is any difference to the 12B4 circuit.

I think the shunt reg if well implemented would have the edge. In my differential ccts I don't recall hearing a massive difference, certainly not worth the stuff around it involved. In low level cct like an RIAA, I'd go CCS/shunt. I had my RIAA's and pre all in the one case, so I used them everywhere.

[kingneb]

I am building a B+ power supply for a powerful tube amp I am making. The design is supposed to supply 360 volts into a 750mA load (720 with the 500 ohm Rload). What I am concerned about is my new RadioShack volt meter, when I set it to AC volts, says there is about 35.35 VAC across Rload. On DC, the meter fluctuates little to none, reading almost a dead on 360.0 volts.

I may have ruined the meter somehow because I accidentally placed the microammeter in parallel with the Rload (moved the switch) and the fuse blew (the fuse does affect the volt meter). I wonder if the AC meter was damaged at all when this happened. My previous meter I did the same thing with years ago. I can be careless, but the meters should be more idiot proof than just a fuse.

Anyway, when I get back to school I will test it with my good Fluke meter, and I have access to a high end HP unit as well.

Is the 35 volts is really there or is the meter giving a fallacious reading? I will try scoping the output tomorrow. If the 35 volts is there, why? I have the following setup:

- 300 volt transformer secondary winding, feeding a full wave bridge. That is about 424 volts minus the ripple and diode drop feeding the regulator, which should be more than enough.

- 100uF filter cap. More capacitance on both the input and/or output does not eliminate the 35 VAC reading.

- The transistor and regulator are heatsinked, as they become warm.

- The schematic of my setup is attached.

- The dummy load I am using is two 1k ohm 225 watt resistors in parallel. They get as hot as a heater, i know.

[SY]

With that kind of current draw, you're going to want a heftier pass device than an IRF820. You may also need to bypass some of the current around the reg using a power resistor to keep the pass transistor dissipation down and keep the 317 farther away from its current limit.

To see what's happening at the output, you need an oscilloscope, otherwise, you're working blind. And that's a pretty dangerous neighborhood to be in without being able to see.

[tomchr]

Sy, in your experience, what's the minimum drop-out voltage these "Maida" regulators will operate at? Also, what causes you to recommend using a 12 V zener rather than the 6.2 V used in the original National App Note?

~Tom

[kingneb]

Putting 12 volts from input to output will raise the voltage differential by a couple volts.

[SY]

Originally Posted by tomchr

Sy, in your experience, what's the minimum drop-out voltage these "Maida" regulators will operate at? Also, what causes you to recommend using a 12 V zener rather than the 6.2 V used in the original National App Note?

It depends on the pass device and the amount of ripple. The dropout is also dependent on the series resistor- note the issue with the rated current and dropout with the originally published schematic. If memory serves, the 317 has a dropout somewhere around 3V, add the Zener voltage minus a Vbe drop (for a bipolar pass device), and then add the raw supply ripple to get an estimate of dropout.

I try not to push dropout, and in tube circuits there's usually volts to burn, so there's no penalty for using a 12V diode; that provides a comfort margin for Nervous Nellies like me.

[tomchr]

SY: Sounds like we're doing the same math. Worst case drop-out for the National LM317 is 2.5 V over temp. Most Maida regulators use a Darlington pair, so there's two Vbe's there. Plus the IR drop through the series resistor on the input of the LM317.

How much current do you generally run through the zener diode? I'm currently looking at 1~2 mA and I think that's a little skimpy. Yet, burning more means burning more power in the series resistor for the zener... Tradeoffs, tradeoffs.

~Tom

[ArtG]

Originally Posted by SY

The number one thing that goes wrong is operator error- a slipped scope probe has killed quite a few pass devices in my lab.

There's great truth in that statement! Why, just this evening I did some "welding" while checking B+ voltages on a tube amp. Managing one probe is reasonable. Placing two while trying to read two separate meters at the same time is challenging!