[jeffreyj]

Quote:

Originally posted by Voodoochile ... Another possibly interesting point to note is that in the LM datasheet, the resistor on the output pin is stated to typically be 220 ohm, regardless of desired output. The LT design notes spec a much higher R value here. ...

To see why this value is so common in the datasheets you need to look a little deeper into the operation of this very clever little regulator. In the LM78xx series, for example, the error amplifier and reference derives their power from Vin and Gnd connections on the device. The LM3x7 regulators have no Gnd pin, so the power to operate the internal circuitry has to come from the voltage differential between Vin and Vo. So, a minimum load must be presented to the output terminal for the regulator to function; the minimum load, in this case, is 5mA (typical; 10mA worst case), and since the minimum output voltage is 1.2V, this requires a 240R resistor. 240R is a standard value for the 1% resistor series but not the 5%, so 220R is a common enough substitute. 120R satisfies the worst case condition, though, so you sometimes see it specified as well. Of course, you can also use a higher value resistor if the regulator is always loaded by at least 5-10mA.

Connecting a 220R resistor from Vo to ground satisfies the quiescent current requirement, but it leaves you with a 1.2V regulator - not terribly useful. Putting another resistor (variable) in series with the 220R one to ground and connecting the ADJ pin to the midpoint allows "ground" to be raised from the perspective of the reference, and voilla, you get an adjustable regulator that can supply a regulated output into the hundreds of volts (as long as the maximum Vin to Vo differential is respected; usually 40V or 60V (HV models)).

 

[Voodoochile]

Squalish: I wouldn't be comfortable running a PS1-type PSU with more than a 1 amp load, as the regs are not up to the task. Certainly the transformer I'm using will not handle more than about 1/3 amp. The reg will take a bit more. But the principle is otherwise very similar. I have not read up on the Gainclones... yet. I have enough on my plate already, but will sometime soon.

JeffreyJ- thanks for the discourse on the resistor. Are any of the pins on the LT1085CT considered a ground? I know there are differences between it and the LM317, being lower dropout, for example. I have used the 317 in this design many times, and it does perform surprisingly well. Not exotic by any means, though.

 

[ppl]

Dreamslacker> Thanks. I notice both LM and LT regs in standard Data sheet configuation make the mids sound slightly hollow compaired to batteries and unregulated supplys but the unregulated supplys are noisy.

using two positive regs as Squalish linked to is a good idea as the negitive regs have a real problem with stability. The resistor i was refering to is not the data sheet 220 ohm but a small 0.5 ohm in series with the output so as to allow lots of output capacitence and still remain stable takes some of that Hollow sound out.

 

[Voodoochile]

Got you now, ppl. Thanks.

Yes, the pair of positive regs looks interesting. I do notice the difference between positive and negative, in both the Linear Tech and 317/337 lines... one of the reasons I like using the trimpots on feedback, as they seldom if even hit the proper output with fixed resistors. A patch, but it works pretty well.

 

[jamont]

Nice work Mark, as usual!

And thanks to Jeffrey for his illuminating discussion of the LM317.

 

[Voodoochile]

Thanks John

.

I think I might either measure this, then swap out for an LT1085CT and remeasure, or maybe even just build another one using the LT reg, but otherwise the same.

 

[Nisbeth]

Nice work Voodoo! BTW, that PSU-family of yours is getting awfully big

/U.

 

[was ist los?]

I've always wondered this, but why is it that you only regulate the posituve rail?

 

[Voodoochile]

I don't think you can regulate a ground*. If you were to regulate a negative rail, then it would be a dual voltage PS. The potential is on the positive rail, and ground merely completes the loop.

At least that's what I heard on TV.

*within the scope of this relatively straightforward design

edit: Thanks Uffe

I have to keep them separated or they start pushing and shoving like little brothers.

 

[Chipko]

The positive rail is regulated against ground. Ground works as the reference potential. Usually you just reference ground to another ground, ie, a tranny has a floating potential but the amp is connected to a source and the two grounds to each other putting them on the same potential. There is no absolute ground. For all practical purposes protective ground is used as a 0-reference.

For audio purposes you want a floating ground since this minimizes the problems with groundloops.

 

[tangent]

You can build a dual-voltage linear supply and ignore the ground rail when hooking it up to a device that expects a single supply voltage. This gives you a regulated positive and negative rails. In most cases, this isn't a very great idea, because you might as well take the next step and hook up the ground rail as well.

 

[Voodoochile]

That is a much better version of what I was thinking, thanks Tangent. What Chipko said plays into this also. You can't really regulate the ground, becase the regulation needs some sort of reference in order to 'know' whether or not it's in regulation.

My verbal skills are beyond horrible.

I'd make an analogy to pole vaulting, just to keep you guys wondering

. Say you run and launch yourself over the bar. but the ground (no pun intended), is moving up and down by several inches. How high have you jumped? You don't really know, because your reference point keeps moving. The ground should serve as your control group, and the height would be your output. I suppose how quickly you accomplished the feat would relate to current, but that's another absurd analogy.

The resistor between the output and the adjust allows the regulator to maintain a voltage reference, or benchmark, and the resistor from adjust to ground determines the output. This resistor is tied to ground, so it allows the regulator to monitor the ground status relative to both the input and output. Is the ground bouncing up and down? Hopefully not too much if you are pole vaulting.

 

[was ist los?]

So would one just hook up the neg rail normally?

 

[Arzela]

If you like, with just a positive rail, you can think of the ground
as negative. It's all relative...

 

[Nisbeth]

Quote:

Originally posted by Voodoochile My verbal skills are beyond horrible. I'd make an analogy to pole vaulting, just to keep you guys wondering . Say you run and launch yourself over the bar. but the ground (no pun intended), is moving up and down by several inches. How high have you jumped? You don't really know, because your reference point keeps moving. The ground should serve as your control group, and the height would be your output. I suppose how quickly you accomplished the feat would relate to current, but that's another absurd analogy. The resistor between the output and the adjust allows the regulator to maintain a voltage reference, or benchmark, and the resistor from adjust to ground determines the output. This resistor is tied to ground, so it allows the regulator to monitor the ground status relative to both the input and output. Is the ground bouncing up and down? Hopefully not too much if you are pole vaulting.

You're doing fine! That's probably the funniest analogy I have ever seen

/U.