TL431 + common base amplifier stability?

[Zero999]

The old TL431 is limited to 36V. For higher voltage circuits it needs to be buffered somehow. A common base amplifier is the most obvious solution but is it stable?

LTSpice says it is stable but I don't trust the model of the TL431.

The extra voltage gain is bound to reduce the phase margin but could this be offset by the fact the gain needs to be high (less negative feedback) and a common base amplifier doesn't have much phase shift? The circuit below has an output voltage of just over 50V, that's a gain of over 20, so it might be stable.

I suppose I'll have to build it and see.

EDIT:
Added .asc file and changed the package to the  PDIP/SOP)/TSSOP so the file can be read, without any bespoke symbols.

[T3sl4co1l]

If you bring in the gain-phase plot for the TL431, you can figure it out.

I would suggest one thing, for two reasons.  An emitter resistor.  First, this regulates the voltage gain of the transistor, so that rather than having transconductance proportional to current (and rather high to begin with, because BJTs are awesome), it's limited by that resistor.  Second, this allows some voltage swing to appear at the TL431, so if nothing else, you can compensate it locally with a C or R+C back to REF.  (Without the emitter resistor, the emitter voltage will be largely the same as the base voltage, less a fixed drop, so the feedback C or R+C will simply feed back no voltage.)

Note that this circuit doesn't have straightforward gain and phase and all that, because there's no independent collector load, it's all shunted back to the same point.  Likewise, base voltage is set by collector voltage, so the transistor looks more like a fixed gain noninverting (not really "common base" anymore) stage.

On the upside, the limited gain will probably correct the ills of the transistor to a frequency much higher than the TL431 understands (i.e., the roll-off for the gain shown will be in the 10s of MHz), and one should be able to make the hand-waving assumption that, as long as feedback gain is less than transistor gain, it will be stable (because, dividing the two out, the TL431 is effectively still running at or above unity gain, and we know it is unity gain stable, so it should still be stable).  That's no guarantee, but it's a start.

BTW, SPICE models of the TL431 are notoriously bad, even as simplified macro models go.  Info:
http://www.audio-perfection.com/voltage-regulators/linear-regulators/realistic-spice-model-for-tl431-stability-noise-impedance-and-performance-simulation-of-tl431-shunt-regulator.html
Helmut is very knowledgeable at LTSpice so, if that's the default model used, you're probably not too bad off.  I would recommend that model, or the "Eugene" model given in the link.

Tim

[Zero999]

Thanks for the reply. It seems like I have a very basic model for the TL431. I do remember downloading it from the Yahoo group but it was some time ago.

I couldn't find a detailed phase/gain plot on the datasheet so it's not something that can be calculated.

I think I'm just going to build this and test the step response with an oscilloscope and signal generator.

[Zero999]

I've tested it and it works. I added a 100R emitter resistor and changed R2 to 47k, in order to reduce the gain and provide more headroom for the voltage drop across the resistor.

It seems to be very stable. The gain of the voltage booster is equal to 1+R1/R2 which is 3.1 and it's inside the feedback loop of the TL431 which has a gain of 19.9 which explains why it's stable, especially as the common base configuration has minimal phase shift.

I also tried reducing R3 to 510R and cranking up the voltage to 65V for a short period of time. It got hot and would probably meltdown if left continuously but it survived a few seconds and was stable.

The output voltage was 50.5V which is within tolerance.

EDIT: Schematic changed to one showing the changes described above.

[schmitt trigger]

An old thread, but I'm glad that I found it.

I had experienced first hand a mild instability in a cascoded TL431 (to emulate a 100 volt zener), and I was unsuccessful in taming it.
I had attempted the usual cures (a compensation cap between the TL431's cathode and its adjustment pin), and had wondered how come it did not appear to behave.

Your following phrase is priceless, and cannot thank you enough. In hindsight it should be obvious!

I would suggest one thing, for two reasons.  An emitter resistor.  First, this regulates the voltage gain of the transistor, so that rather than having transconductance proportional to current (and rather high to begin with, because BJTs are awesome), it's limited by that resistor.  Second, this allows some voltage swing to appear at the TL431, so if nothing else, you can compensate it locally with a C or R+C back to REF.

[Yansi]

My limited experience in building pretty good fixed voltage HV supplies (100-300V) via TL431 in this configuration  tells that this indeed is pretty stable configuration.

//Edit: That emitter resistor, I am not sure if I have used that or not, but I think I have not, yet it still was stable - or at least stable enough it never oscillated

[schmitt trigger]

I have no doubt that you have built stable high voltage references. From other threads of yours, you appear to have significant knowledge with HV circuits (with HV simplistically defined as anything larger than 30 volts).

But as I mentioned earlier, I was unsuccessful in my previous attempt..... So the trick mentioned above is something that I might try in the future.
If the need arises.

[Yansi]

There also may be significant parameter variation across different specimens and vendors of TL431, so I understand the concern about the sim model not being "correct enough". I just may have been more lucky with picking "a good one". I am nowhere near any expert HV supply designer, I just love challenge and linear HV supplies are quite it. And still very useful tools for the shop.

Do you have any thread here with a schematic of your unsuccessful build, so I could have a peek there?

[001]

My limited experience in building pretty good fixed voltage HV supplies (100-300V) via TL431 in this configuration  tells that this indeed is pretty stable configuration.

//Edit: That emitter resistor, I am not sure if I have used that or not, but I think I have not, yet it still was stable - or at least stable enough it never oscillated

Can You say, what is the best way to replace VR tube with TL431?

[Zero999]

An old thread, but I'm glad that I found it.

I had experienced first hand a mild instability in a cascoded TL431 (to emulate a 100 volt zener), and I was unsuccessful in taming it.
I had attempted the usual cures (a compensation cap between the TL431's cathode and its adjustment pin), and had wondered how come it did not appear to behave.

Your following phrase is priceless, and cannot thank you enough. In hindsight it should be obvious!

I would suggest one thing, for two reasons.  An emitter resistor.  First, this regulates the voltage gain of the transistor, so that rather than having transconductance proportional to current (and rather high to begin with, because BJTs are awesome), it's limited by that resistor.  Second, this allows some voltage swing to appear at the TL431, so if nothing else, you can compensate it locally with a C or R+C back to REF.

Interesting. I did some more tests without the emitter resistor and found it to be stable. I think keeping the voltage across the TL431 reasonably low also helps, because it means less negative feedback can be used for the BJT. I imagine there's a greater risk of oscillation with larger BJTs or several stacked, because there will be more phase shift. The BC546 on its own should be much faster than the TL431, just not cause any meaningful phase shift at the frequencies it cares about.

What BJT did you use? Was it a big, high current one?

My limited experience in building pretty good fixed voltage HV supplies (100-300V) via TL431 in this configuration  tells that this indeed is pretty stable configuration.

//Edit: That emitter resistor, I am not sure if I have used that or not, but I think I have not, yet it still was stable - or at least stable enough it never oscillated

Can You say, what is the best way to replace VR tube with TL431?

What voltage regulator do you need to replace?

[001]

Can You say, what is the best way to replace VR tube with TL431?

What voltage regulator do you need to replace?

Classic
5651 http://www.r-type.org/pdfs/5651.pdf
0A2 http://www.mif.pg.gda.pl/homepages/frank/sheets/035/0/0A2.pdf

[Yansi]

I have always used small-medium signal fast NPNs from TV video amplifiers for the cascode.

Such as BF259, SF359, or 2SD1264 may work sufficiently.  Do not use large power BJTs, such as BU508, they are slow as ... just too slow.

[floobydust]

Similar approach with 280V 0.1A and TL431, no zener. Schematic is from http://www.next-tube.com
pdf has explanations and measurements of the specs 0.35% line, 0.5% load regulation, max 420VDC input.

[T3sl4co1l]

I have no doubt that you have built stable high voltage references. From other threads of yours, you appear to have significant knowledge with HV circuits (with HV simplistically defined as anything larger than 30 volts).

But as I mentioned earlier, I was unsuccessful in my previous attempt..... So the trick mentioned above is something that I might try in the future.
If the need arises.

Try putting a resistor in series with the emitter, and a small R+C across the TL431 (R6, R7, C2 in flooby's attachment above).

Note that this configuration is only rated up to the 100mA the TL431 can draw, and down to the 1mA it's specified to regulate tightly to; this is much wider than a VR tube so it's plenty fine for that application.

The cascode base voltage needs to be set to limit TL431 dissipation; at say 50mA it should dissipate less than 625mW, but preferably a lot less, like 300 or 150mW.  That's 3-6V.  We can allow more cascode voltage if a resistor is used, sized to reach this point anyway; so, if it drops 6V at 50mA, it's a 120 ohm resistor, and the cascode can be 12V.  (The cascode can also be set by a zener diode instead of a resistor divider, which is convenient if you want to make a variable regulator.)

The transistor needs to be rated for power dissipation and SOA, which should be fine for most options.  BJTs: prefer audio/ linear / driver / power output types, with wide SOA.  150V at 30mA is 5W, so a well heatsinked TO-126 will do, MJE340 for example.  MOSFETs: check that it has a wide SOA.

Prefer smaller devices (lower capacitance), which will have less impact on performance and less need for compensation.  Check compensation with a source or load step test.

If the cascode divider has a low impedance (zener or cap-bypassed), consider a ferrite bead or series resistor to the base/gate terminal, to prevent oscillation.

Tim

[001]

Such as BF259, SF359, or 2SD1264 may work sufficiently.  Do not use large power BJTs, such as BU508, they are slow as ... just too slow.

Thank You!

That is "slow"? Is 4MHz not enough for this circuit? 

[Zero999]

Where did you get 4MHz from?

What range of currents/voltages do you plan to operate it over? If it's fairly narrow, a resistor can be used.

For example:

V = 85V
I = 2mA to 3mA

Cheat and use the calculator linked below to calculate R1 and R2:
https://www.random-science-tools.com/electronics/divider.htm
V_IN = 85V
V_OUT = V_REF = 2.495V
Preferred value series = E24

R1 = 430k
R2 = 13k

I_R1 = I_R2
I_R3 = I1- I_R1

I_R1 = V/(R1+R2) = 85/(430k+13k) = 85/443k = 0.192mA

When I1 = 2mA
I_R3 = 2mA-0.192mA = 1.81mA

When I1 = 3mA
I_R3 = 3mA-0.192mA = 2.81mA

The minimum acceptable voltage drop across the TL431 is 2.5V and the maximum is 36V. Therefore the voltage across R3 could vary from 85 - 36 = 49V to 85-2.5 = 82.5V and the TL431 will be fine.

If we set R3 so it drops 82.5V at 2.81mA.

R = V/I = 82.5/2.81mA = 29.36k.
Choose 29k, which can easily be made with a 16k and 13k in series.

Calculate the voltage on the TL431, when I = 1.81mA
V = 85 - 29k*1.81mA = 32.5V

Calculate the voltage on the TL431, when I = 2.81mA, to check the rounding hasn't caused any problems.
V = 85 - 29k*2.81mA = 3.5V

EDIT: The voltage at U1's anode, as the current is swept from 2mA to 3mA.

[001]

Thanx fo Your equations!

Is it schematic usable as is? Need I add some diodes to protect it at on/off?
What currenrt range practically achived?  0A2 tube is ok up to 30mA

Where did you get 4MHz from?

From BU508 datasheet. Why Yansi named it "slow"?

[Zero999]

Thanx fo Your equations!

Is it schematic usable as is? Need I add some diodes to protect it at on/off?
What currenrt range practically achived?  0A2 tube is ok up to 30mA

Whether it's usable, depends on your application. I had the 5651 in mind when I did it, with the idea the current will be nominally 2.5mA, 2mA minimum, 3mA maximum.
http://www.r-type.org/pdfs/5651.pdf

The circuit only works over fairly narrow current ranges, because the voltage across the resistor varies, as the current varies. If the current is increased, beyond what it's designed for, a greater voltage will be dropped across the resistor, causing the voltage across the TL431 to fall below 2.5V and at lower currents, the voltage across the TL431 will increase to above 36V, which will damage it.

A zener diode could be used to get a wider current range. You might ask, why not simply use a zener diode in the first place? Well you could, but the TL431 has a much more stable output voltage, which varies less, as the current changes. Another possibility is a circuit called a V_BE multiplier, which again uses a BJT, but it's less stable than a zener.

[schmitt trigger]

Thanks everyone for providing  so many valuable suggestions.

As I mentioned earlier, this was a while back and did not really kept records of exactly what cascode transistor I employed back then, other than it was a 60 volt, TO220 in a Radio Shack blister package.

I know, I know....... A generic power transistor from Radio Shack.  Embarrassing.
Edit, or perhaps it was 80 volt transistor? Can’t remember.

[T3sl4co1l]

I know, I know....... A generic power transistor from Radio Shack.  Embarrassing.
Edit, or perhaps it was 80 volt transistor? Can’t remember.

Probably TIP31C or something like that.

Tim

[schmitt trigger]

You may be right about the TIP31C.

I recall that back then those transistors were very popular medium power, low cost devices.