Thinking about it, 0.4 V seems like an awfully large error for the post #2 circuit. Are you sure the circuit is not oscillating?

ak

 

@upand_at_them I assume you have to subscribe to read the articles?

 

sure the circuit is not oscillating?

The output is clean.

 

Why so much capacitance before the regulators?

More F = less ripple. A quickie approximation for p-p ripple voltage is EC=It, or E = (i x t) / C. Units are volts, amps, seconds, farads.

You left off the typical diode protection and smoothing tantalum caps after the regulators?

This is not my schematic. I found it in a search to illustrate the concept. Output filtering/transient response components, high-frequency bypass, reverse-voltage protection - all recommended for any power supply design. I updated the post.

Note: some active regulator components have a maximum output capacitance specification to prevent oscillation.

ak

 

More F = less ripple.

Yes, but the regulator does that also... I'm getting a good clean output now with just the 1000uFs on each side. The output LEDs and switches were a nice touch also. I like to overall design and kinda follow how the Op-Amp does the tracking. Nice addition. Looks like I might have build version 2 soon to incorporate the tracking.

EDIT: In fact, I may delay my build to at least breadboard the tracking Op-Amp and consider adding it to the build. Now that I have the 2 tiny PSUs from eBay I at least have a dual voltage supply available even if not very many mAs.

 

@upand_at_them I assume you have to subscribe to read the articles?

Yeah, you can purchase back issues. And they may have already gone digital by 2012. I don't remember when my subscription ran out, if I still have that year.

 

@AnalogKid OK, I don't have an LM301. I do have the suggested replacement OP07 (if suitable) which has even a bit higher supply voltage. Problem is the diagram is for a 12V secondary and I am using a 28V secondary XFMR and after the rectifier getting 22V. So I will have to drop the Op-Amp supply voltage to under 18V. A Zener?

EDIT: Nope, I need the opamp to be able to handle the ~22V input to track... The OP07 maxes out @ 14V so I need something that can take the 22V input. The 301 maxes @ 22V supply but can handle 30V input. I need something similar that can handle 24V supply or higher depending on swing to rail.

 

EDIT: Nope, I need the opamp to be able to handle the ~22V input to track... The OP07 maxes out @ 14V so I need something that can take the 22V input.

I haven't more than glanced at the schematics. Do you really need the opamp supplies to be symmetrical?

Tektronix did this in their PS503A triple power supply:

Both opamps are uA741. They used mostly uA741's for opamps in many of their designs, using better only when necessary.

 

27V supply into a 741? It is only rated for 18V and 15V input?

EDIT: Ah, the uA version is higher but still shows 22V max for supply but does go 30V on input.

And no they are not symmetrical at all the neg. side is grounded since it is following a positive voltage only.

 

27V supply into a 741? It is only rated for 18V and 15V input?

EDIT: Ah, the uA version is higher but still shows 22V max for supply but does go 30V on input.

And no they are not symmetrical at all the neg. side is grounded since it is following a positive voltage only.

Might have been a screened part. My hardcopy says the manufacturer part number is MICROA741CP and TI is the manufacturer.

 

I do have some uA741 and LM301 on order from china but will be quite a while before I get them and then test them. My last order of uA741s was bad...

 

They're not screened parts.

Here's the TI datasheet fragment:

What this really means is that the absolute maximum difference between the supply rails is 36V. You could have V+ be 200V and V- be 180V and the opamp is still within it's allowed supply voltage range.

I've seen opamp circuits with an output swing of 120V, but they adjusted opamp supply voltage dynamically so it never exceeded the allowed differential voltage.

 

TI's MA741CP is also showing 22V supply and 30V input.

EDIT: Ah, so with the neg. grounded the pos supply could go 36V for a max 18V input opamp?

 

TI's MA741CP is also showing 22V supply and 30V input.

EDIT: Ah, so with the neg. grounded the pos supply could go 36V for a max 18V input opamp?

Not sure what you're getting at. Typical rails for uA741 are +/- 15V, giving a 30V supply differential. Absolute maximum supply differential is 36V and Tek is operating at 32V. I don't see any problems with the design.

 

So the absoulute max +supply voltage with the negative grounded is +44V?

 

Got the opamp voltage follower sketched out and will try breadboarding it with a 741 later today.

 

View attachment 220216
So the absoulute max +supply voltage with the negative grounded is +44V?

Yes, but you never operate a part at it's absolute maximum rating because the devices aren't guaranteed to survive that condition. Even if they did, 44.001V would likely kill it.

 

27V supply into a 741? It is only rated for 18V and 15V input?

Yes. Look more closely at what Tek is doing. The algebraic difference between the power rails of only 32 V for each opamp. Because of the 741's headroom requirements, you need at least 3 V beyond the desired output voltage range. So, for the positive output, if you want it to go down to true zero volts, the opamp negative rails must be beyond that. That is why the positive output opamp negative rail is -5 V, and the negative output opamp positive rail is +5 V.

There is nothing in the datasheet for the 741 or almost any other general purpose opamp that requires the supplies to be balanced. Running a 741 on +36 V and GND is just fine.

ak

 

K, great! Got it, was being thrown when they said +/-22V max as I assumed that was from gnd. With it being the total span that leaves a lot of headroom here since the rectifier maxes ~24V and one side to gnd.

 

Well, I was having a EUREKA moment. Dialed it in for 5.07V and the negative followed at -5.0678. But when I started dialing the plus up the negative quit following ~-7.4V. It was tracking perfectly then just stopped around -7.4? I left the output protection diodes and the 2.2uF tantalum output smoothers on the circuit and had it stripped down to the rectified feed and regulators. Somehow it just quit following.