At a fork in the road with PSU design

Help Support GroupDIY:

NewYorkDave

Well-known member
Joined
Jun 4, 2004
Messages
4,378
Location
New York (Hudson Valley)
For a designer, one nice thing about a forum like this is that even when you're capable of weighing your options and deciding on your own, it can be enlightening to present the choice to others and see which route they would choose.

I have a tube circuit that can be broken down into three major sections. The first requires approx. +130V regulated at 0 to 14mA. The second section needs +130 to +150V at 6mA and the third requires +130V to +150V at 18 to 20mA. (The B+ for sections 2 and 3 must be well-filtered, but not necessarily regulated). I intend to provide the regulated supply with a TL783. My raw DC input is about 175V, capable of up to 100mA.

My first inclination was to split the supply after the input filter and feed sections 2 and 3 off a simple R-C filter, and section one off the regulator. Since section 1's current draw varies, the voltage divider resistors on the 783 must be sized for a fairly large standing current (15mA, IIRC).

Then, it occurred to me, why waste that 15mA? If I fed the entire circuit off the regulated supply, there would always be ample current (24mA minimum) flowing from the 783's output pin to assure proper regulation--unless all the tubes were removed, of course. And one less filter cap would be needed. On the downside, the regulator would have to dissipate more heat at maximum load (40mA max vs. 14mA max).Sections 1 and 3 are push-pull, so along with the low output impedance of the power supply, cross-coupling through the supply seems pretty much a non-issue.

What do you think?
 

Svart

Well-known member
Joined
Jun 4, 2004
Messages
5,134
Location
Atlanta GA USA
It's always these kinds of tradeoffs in the manufacturing world. What you need to ask yourself is, what will each one of these designs do to the sound?

If they don't *change*(yes i know everything can change the sound) then it's always wiser to go with the more simple route. It's cheaper, easier and there are fewer things to go bad.

I sometimes find it easier to start from what you want the end product to be like and work backwards, especially if you have a very specific intended outcome. If the outcome is less specific and can stand to be varied slightly then it's easier to start from the basis, which is usually the powersupply.

This maybe just because I have worked in production environments or that I am a minimalist at heart.

Another thing you mentioned that is interesting is the varied HVs for the stages, since the least common denominator is 130v then do you plan on giving all the stages the same HV or do you intend to give the other stages a different HV? What would you like to do? Would it change the sound to do it differently? Just some of the same questions I would ask.

So it seems that I have eluded the direct question of what I would do so I will answer it now, I just wanted to pose some rhetorical questions to help you find your own answer.

I would go with one HV rail of 130v comprised of rectified line(~170v) with good decoupling. Now we get to MY crossroads, I would likely try the 783, but I would also like to try a discrete mosfet solution too. Either one is likely to work fine if setup correctly to the circuit. With the mosfet, you could likely use a part large enough to easily deal with most any current and dissipation you needed. You could also try an IGBT but they do cost a bit more than a MOSfet.

Just a thought.
 

analag

Well-known member
Joined
Apr 23, 2005
Messages
1,718
Location
Mars
My answer 783 regulated @ 150V then zener/MOSFet @ 140V then the same zener/ MOSFet combo @ 130V.

analag
 

pstamler

Well-known member
Joined
Jan 24, 2005
Messages
1,509
Location
St. Louis, MO, USA
I think I'd go someplace similar to what Analag is suggesting, but I'd probably use Darlington pairs of bipolar transistors rather than mosfets.

The other thing to look at: how much voltage will be across the TL783 (including during turn-on and turn-off), if the line voltage happens to be 10% high?

Peace,
Paul
 

NewYorkDave

Well-known member
Joined
Jun 4, 2004
Messages
4,378
Location
New York (Hudson Valley)
...And here's yet another idea, simple and relatively bulletproof.

Image

The regulation won't be as stiff as it would with an active solution, but it may be good enough for the job at hand.
 

analag

Well-known member
Joined
Apr 23, 2005
Messages
1,718
Location
Mars
[quote author="rafafredd"]I smell vari MOOOO...[/quote]

Haha....you could be wrong, but let me share my experience with the 660 type I have. I hold the 130V supply with an iron fist.

analag
 

NewYorkDave

Well-known member
Joined
Jun 4, 2004
Messages
4,378
Location
New York (Hudson Valley)
Yes, it's for a Very-Moo. The circuit is still under development... But I will say that all you've heard about the virtues of regulated B+ for the gain control stage is totally borne out by my experiments.
 

Gus

Well-known member
Joined
Jun 3, 2004
Messages
4,795
Location
n
Dave

How much current(and power) in the zener before the tubes warm up enought to shunt the current their way. At power up the zener will have more than 4ma I would think.

I like shunt supplies. I like to have a good amount of the circuit dynamic current flowing in the zener(s).

With some stuff say the circuit uses 1ma I will often set the zener current to 2ma or more with the load disconnected.

Maybe a small resistor between the diodes and first cap to help a little with Si switching noise??
 

NewYorkDave

Well-known member
Joined
Jun 4, 2004
Messages
4,378
Location
New York (Hudson Valley)
Yeah, the zeners have to carry the full value of whatever current it takes to drop the B+ to 130V, during any no-load condition, and must be rated appropriately. For good regulation, several milliamps should be flowing through the zeners at all times. The best value depends on the zener, but approx. 5mA is a common rule of thumb.

The good thing about a shunt-regulated supply, if you're putting it at the end of a chain of R-C filters, is that it's going to draw a nearly constant current even as the load current fluctuates (under conditions of normal operation--not counting, say, an accidental short!).
 

solder_city

Well-known member
Joined
Oct 18, 2004
Messages
325
Location
new jersey, USA!
plus theyre...

studly.

:cool:

three legged critters and high voltage dont seem to mix, despite what ratings might say. one little mishap and poof! personally id go with the zeners, or even a gas regulator tube (remember those)!
 

analag

Well-known member
Joined
Apr 23, 2005
Messages
1,718
Location
Mars
[quote author="NewYorkDave"]Yes, it's for a Very-Moo. The circuit is still under development... But I will say that all you've heard about the virtues of regulated B+ for the gain control stage is totally borne out by my experiments.[/quote]

My VU meter sits the usual way in the B+ line...and she swings her ass off.
 

pstamler

Well-known member
Joined
Jan 24, 2005
Messages
1,509
Location
St. Louis, MO, USA
During turn-on (or when a tube decides to go belly-up) each zener will be dissipating about 0.6W. A 1W zener will work but the margin's a little tight. I'd go for 5W.

Peace,
Paul
 

Larrchild

Well-known member
Joined
Jul 15, 2005
Messages
801
Location
Fort Lauderdale, Florida
I used a pair of 100v 5 watt zeners as a shunt reg for my 407a mic pre. But thats just 4 tube sections.

For a stereo mu unit, of course you need a pass device. and for best seperation-intermod, 2 pass devices.
 
Top