Anode voltage (270VDC) regulator circuit

Hello everybody

I am working on a valve guitar preamp and have some trouble with getting anode voltage supply with no ripple. I do not care whether the voltage will be 260 or 280 volts or will drift slightly with temperature, just don't want 100Hz frequency in it.

At first, I made a simple regulator using a series transistor with a Zener diode in its base (a common collector configuration). This circuit works fairly well, but due to high input voltage swing (approx. 60Vpp) and finite dynamic resistance of a Zener diode (which was biased from the unregulated part) the ripple in the output voltage was about 300mV. Though it's not too big compared to 270V output, I am trying to improve this result. Also because high voltage Zeners have poor parameters and get heated up quite a bit.

I came up with such a circuit:

SIMULATION

I employed an op-amp as an error amplifier, with 5V6 Zener as a reference and a resistor divider as a negative feedback, so the op-amp will survive the voltages present. The signal from transistor gets amplified by a common base transistor voltage amplifier, so it can drive the power transistor's base. The transistors I would like to use are MJE-13003, op-amp TL071. My supply is a step-up transformer (1020:965 turn ratio) powered from 230VAC. I need at most 50mA of output current.

I would like to get some advice wether this circuit will work in practice, perhaps some comments on common base amp resistor values.
Thanks in advance, Marcin.

Your link is not clear
what do u want us to see on that page.
send exact diagram , and how r u getting so much swing after FULL WAVE RECTIFICATION what is the value of the smoothing capacitor ? I*t=v*c I=current..t=1/100hz calculate c value and use more than double value that u got and least 300v dc
google hv pwr supply, and all that explanation op.. error better post the diagram.

https://w5jgv.com/4-400A_600M_amplifier/POWER.pdf

https://www.qsl.net/k3pd/chap14.pdf
see chapter 14

The big error on the simulation is the output tap-off voltage with its 2M resistor, the negative input of the error amp must be returned to the output HT line via a capacitor so there is no ripple reduction at all. You will also have to scatter some diode around this area so the initial changing for this capacitor does not overload the opamps input, like wise at switch off the residual charge on the capacitor does not reverse bias the opamp.
If you have > 230 V RMS, and a bridge rectifier with a reservoir capacitor of 470 MF, you will have about 370 Volts DC, i.e. about 140V DC you can loose, at 50mA = 2.8 K, so if you have two RC sections with a 1.4 K resistor and the biggest caps you can find it should work OK and be bullet proof.
Frank

I would foresee a local compensation network around the opamp (a simple feedback capacitor may be enough).
Also I'd prefer a PNP on the high side - it should provides better rejection.

Thank you all for your help.

@dselec Well, in the initial version of the regulator I had only a 47uF capacitor- thus the swing. I also googled with the keywords you adviced and found dome other regulator:

I can see that the double transistors in the reference voltage network are used only because a single one wouldn't survive the voltage. I could replace them with only one because my ones have higher voltage rating. Another advantage of this circuit is that it doesn't have an op-amp, so there's nothing that could die frim overvoltage during transient conditions. I have to think about this one.

@chuckey I don't quite understand where should I place that capacitor you mentioned and why doesn't this circuit cut the ripple out without it. And as for the RC filters- they're OK, though they take a lot of place and I'd just prefer the voltage to be guaranted at a set level.

@dave9000 What's the capacitor for? Do you mean it like in the integrating circuit? And why would a PNP transistor provide a better rejection? By the way, I can't think of the way how should I connect a PNP in a current- emitting voltage regulator.

pietruch said:

@dave9000 What's the capacitor for? Do you mean it like in the integrating circuit? And why would a PNP transistor provide a better rejection? By the way, I can't think of the way how should I connect a PNP in a current- emitting voltage regulator.

Click to expand...

The rejection is better (in principle) because both transistors have high output impedance. Of course the ripple due to the load may get worse.

Your regulator without load takes about 13ma and thats a lot.
input cap taking 360vdc could be 2x47uf, output 1x 4.7uf is better.
u may have a faulty input bridge rectifier , so disconnect every thing and check dc and ac value at the bridge output and post the results.
then attach the 47uf capp to the bridge and post result , next acc 47km high wattage resistor as load to the cap and post result.

I checked the rectifier with an oscilloscope, appears to be working fine. A nice abs(sin) wave with approx. 320V in peak with a 22kOhm load, no cap. Then, after adding the 470uF cap VDC was 309V and the AC ripple was only 300mV with the same load! Maybe the cap I used before was broken? I can't remember which one was it because it got mixed up with others in a box. BTW- it was 47uF, now I used 470. So, if the previous, ultra simple schematic:

worked with 100mV output ripple when supplied with voltage swinging 60V, maybe it's sufficent? I don't mind 200V+ zener terrible temperature coefficent. I'll assemble it again tomorrow with this new cap and post the results.

pietruch, you don't say whether or not you have a choke following the rectifier. If not, I would suggest that you try one. It used to be standard procedure when valves were in common use.

Good well done keep it simple.
My advice use 3 zeners in series and add 47k parallel on each zener .
don't worry about temp... even +/- 10v drift output will do.
have u considered using only zeners and one resistor ?
does the supply +vcc change a lot because the mains supply swings ?

Hello.
I've even got a 0,75H choke taken out from an old half-valve half-semiconductor TV. I wouldn't want to use it though- I'd prefer to add another cap. Choke is big, heavy and interfere with magnetic field. I know that they used to experimentally optimize positions of chokes and supply/speaker transformers to minimize the 50HZ hum. I'd just prefer to skip this part .

As for the only zener stabiliser- If I wanted to get say... 270VDC with 60mA max output, I would have to use zeners that would alltogether dissipate over 16W power (when no current would be drawn from the supply), whereas adding only one transistor makes the regulator circuit draw say... 540mV (2mA for 270V zeners bias) quiescent power and only about 40V (say 310 supply - 270 output) get dissipaded on a transistor. I'll assemble this circuit in the evening and post the results.

Have u calculated the max current u will be taking from regulator ?
after all is ok and working i advice adding a resistor in series to collector to lower the vce * ic dissipation on transistor just to be on the safe side if transistor tor shorts and yes i would add zeners say about 10v more at output more than output volts and afuse series to collector.
good luck

So... The measurements
Concitions- load = 22kOhm
Ripple Vpp at a capacitor: approx. 450 mV
Ripple Vpp at a load: approx. 30mV (but it's 50 Hz not 100Hz!)
Current through 22k load: 12,7mA
Transistor dissipation: 12,7mA * 28V_Uce_avg = 355 mW (I used two MJE 13003 in a darlington configuration, due to small hfe).

I think that the ripple on load come exactly from the transformer. Voltage on Zener diodes doesn't follow ripple on capacitor, so it's not due to too smal dynamic resistance of them. It looks exactly like the interference that my scope takes when placed near the transformer. I'll try later to shield them or get them away and post what happens .