Super Regulator

When I visit the parts search website , and when I type "2222A" into its search box, octopart returns some surface mount transistors in the SOT-323 package and others in the SOT-23 package. MouserUSA has > 200K of them in stock and on the shelf, ready to dispatch.

Similarly, when I visit and surf to their Discrete Semiconductors -> Transistors -> Bipolar Transistors chart, and then type "3906" into the "Search within results" box, I get lots of different options in surface mount packages. MouserUSA has > 1.6 million of them in stock and on the shelf, ready to dispatch.

A quick comparison of the surface mount transistor's datasheet, versus the thru hole transistor's datasheet, confirms or refutes the hypothesis that they are interchangeable (except for power dissipation!).

Thanks Mark for pointing out what turns out to be the obvious…
Hello, I bought PCB, and I don't know which capacitors use, on image I saw Panasonic FC, but I don't like Panasonic FC - because of lack of transparency, somebody suggested that capacitors should have low leakage current, any ideas? Nichicon KZ or something better? Something from Wurth Elektronik maybe? Is low ESR advantage there?
Joined 2011
Paid Member
I very much like the Nichicon URZ series of aluminum electrolytic capacitors ( link) because they specify a super wide temperature range and also because they are physically smaller than many other caps at the same capacitance & voltage. You will have to measure ESR yourself, it does not appear on the Nichicon URZ datasheet.

Occasionally I'll buy the Nichicon UVY as well (another link) -- sometimes they are even smaller.
Joined 2019
Paid Member
This seems to be an ongoing back and forth discussion with no resolution.
Just for a little focus (hopefully) how low an ESR number would be a problem? Is there a number... a dividing line?
After years of observing this debate the impression is that for the most part it doesn't matter.
But if it did...
Hope this helps.
Joined 2011
Paid Member
Maybe start by reading/studying the original magazine articles by Walt Jung "Regulators For High Performance Audio, part 1 and part2". That's where he unveiled the Super Regulator and exhibited it to the world. Check whether Walt mentions ESR, and provides guidelines saying (a) how much is too much; and also (b) how little is too little.
Joined 2019
Paid Member
I could do that (again). And I will again when it comes time to stuff my boards. But I summarized my recollections and impressions... it doesn't much matter.
But what I'm suggesting is that the cautions expressed so often here could just as easily be quantified so that it's clear. At least more helpful to the asker than "Walt has warned against low ESR capacitors". What constitutes a "low ESR" exactly?
How long is that piece of string? Here be dragons.


That said I agree with reading the texts... all the way back to Mike Sulzer.
Two more questions about alternative component experiments:

1. If using ADA4625-1: is it important to solder the exposed pad to GND?

2. On a scale from 1 to 10, how important is it to add the clamping diodes on the opamp’s inputs to the board? Or: in what situation might voltage spikes occur that would require the clamping diodes for protection of the opamp?
1. If using ADA4625-1: is it important to solder the exposed pad to GND?

Before you solder a new alternative part "in", make sure the SR works with it. There are some opamps which won't start, although from reading the datasheet it would appear this is not the case for a low noise JFET opamp like the ADA4625.

If I recall correctly, the DIYAUDIO Store PCB did not have the pad designed in, so you'll have to roll your own. If, then, you are making your own PCB's put a copper layer on the top and bottom for the outline on page 35 of the datasheet, these are connected with vias. Se also the discussion of thermal management.
Thank you. My question was prompted by the datasheet's suggestion for the exposed pad:


As it says the pad may be left floating, I wondered whether there is any pressing reason in this circuit to connect it to V+, V-, or GND plane. Having read further, it seems the answer is:


So the question is whether this circuit makes the kind of demands on the opamp that causes heat to become an issue.
I'm trying to design SR with higher voltage output i.e 50-55VDC, using for front end stage in class AB amp. Why we limit SR only for line stage, pre-amp ect...?
So, thinking about using high voltage op-amp like OPA445 or using SR's popular AD825 (simple regulator with single transistor + 1 zener diode to keep Vsupply for op-amp below 40VDC)
What way is better?
Firstly i only see that is money. High voltage op-amp quite expensive, few tens dollars per each.

Anyone may share to me some advices?

Joined 2011
Paid Member
You and your circuit design consultants will need to make some calculations, to ensure that each component of the SR is operated within its datasheet specification limits, and also to ensure that maximum junction temperature limits are not exceeded.

I recommend you start by calculating the power dissipation of series pass transistors "Q1" and "Q3" in the Super Regulator schematic. Do they require heatsinking? If so, what thermal resistance rating of your heatsink is needed, to keep transistor junction temperature well below the Absolute Max specification on the datasheet?

Next, I recommend you calculate the required base currents of Q1 and Q3 in the worst case (lowest possible beta, highest possible regulated output current).

Next, I recommend you calculate the current that your opamps are required to sink/source {flowing through R3 and R10}. Are the currents comfortably small? Uncomfortably large?

Then, working with your team, calculate the margin-of-safety for the rest of the components on the PCB -- besides the two opamps and the two series pass transistors.



  • SR_schem.png
    24.2 KB · Views: 89
Hi Mark,

Thanks for your advice.

I already read your comments before about load current/Hfe of pass transistor. Yes, if you run load current at ampere, that is trouble. D44H11 is low gain.

I design only for front-end stage including LTP, VAS/TIS which draw max 50mA and it also don't require too high performance. Just add more 1 transistor at output op-amp to isolate and increase PSRR (at least LTSpice tell me that).

When playing with different components in LTspice, I ran into a problem. Output voltage is set to 30V. I'm not trying any complicated analysis. Just a transient spice directive to be able to verify voltages and currents at different places in the circuit. Raw DC input is 34V with a 100mV / 10ms sawtooth ripple applied on it to be able to see if the circuit operates properly.

My simulation works as expected with the schematic and parts listed in the store. When I swap out the AD825 for the ADA4625 the simulation won't run - I see this:

Screenshot 2024-03-12 at 17.30.04.png

LTspice keeps running a 'pseudo-transient' analysis and never gets past that (waveform viewer never appears if I wait long enough for the algorithm to complete and shows garbage data if I escape out of the loop). I googled round for this and it may be that LTspice can't find a DC point to operate after running several algorithms. But that's me just guessing really and don't properly understand what that means anyway.

With the AD825 in the schematic the waveform viewer pops up instantly when I run the simulation and shows data that makes sense. By the way: I added the AD825 component to LTspice after downloading the corresponding .subckt from the AD web site; the ADA4625 spice model is native to LTspice.

I could ask on an LTspice forum, but usually when I can't LTspice to do something I want it's because the circuit under simulation has errors. And that would bring me back right here...

Would anyone be able to help me understand what's going on here?