I'm making a regulated 150v power supply for an OTL headphone amplifier. Supply current at idle is around 100mA and can go up to 200mA (class A).
I'm making a zener follower circuit and it works but not for long. Somewhere along, the TIP122 Darlington died causing unregulated 180V supply to show up at the emittor (collector-emittor shorted).
Schematic is based on this but i made some changes i described below:
Voltage drop resistor value is 2K2. TIP122 darlington. Zener is series of 36v and 120V zener. Input cap is 220uF/200v and the only difference with the schematic above is i put 100uF capacitor across the zener (so if you remove the zener, it's a capacitance multiplier circuit).
1. Why did my TIP122 die? It could be that i fiddled and somehow shorted the transistor by mistake (so nothing wrong with the schematic).. i just want to make sure before i put a new darlington. Here's the datasheet.
2. Does zener really have bad regulation? 36 + 120 = 156 v but i'm really getting around 168-170V at the resistor-zener junction. This is why i fiddled with the circuit and may have killed the transistor.
I'm making a zener follower circuit and it works but not for long. Somewhere along, the TIP122 Darlington died causing unregulated 180V supply to show up at the emittor (collector-emittor shorted).
Schematic is based on this but i made some changes i described below:
An externally hosted image should be here but it was not working when we last tested it.
Voltage drop resistor value is 2K2. TIP122 darlington. Zener is series of 36v and 120V zener. Input cap is 220uF/200v and the only difference with the schematic above is i put 100uF capacitor across the zener (so if you remove the zener, it's a capacitance multiplier circuit).
1. Why did my TIP122 die? It could be that i fiddled and somehow shorted the transistor by mistake (so nothing wrong with the schematic).. i just want to make sure before i put a new darlington. Here's the datasheet.
2. Does zener really have bad regulation? 36 + 120 = 156 v but i'm really getting around 168-170V at the resistor-zener junction. This is why i fiddled with the circuit and may have killed the transistor.
Hi,
In theory the schematic ought to work. But you need to take in account, choosing the transistor, that the load might have a capacitive component. This might cause the Vce to rise above the specified value for the TIP122 at startup (damaging the transistor more and more until it fails). I'd try a 200V type...
That also gives you the opportunity to bypass the zenerdiode with a capacitor, reducing noise.
Cheers
In theory the schematic ought to work. But you need to take in account, choosing the transistor, that the load might have a capacitive component. This might cause the Vce to rise above the specified value for the TIP122 at startup (damaging the transistor more and more until it fails). I'd try a 200V type...
That also gives you the opportunity to bypass the zenerdiode with a capacitor, reducing noise.
Cheers
Last edited:
I see.. this is because it takes time to charge the capacitor? Meaning at start up, base voltage will be close to zero and rising until it reaches the zener breakdown? So.. my options would be to remove the capacitor or change the darlington with higher Vcbo, right??
Vcbo TIP122 is only 100V.. i guess this is the rating that i stepped over?
Keep the capacitor. Go with higher Vcbo.
You got it.
A smaller capacitor will rise in voltage a lot faster too. Think about that. You can still put a filter capacitor on the output; but take its surge into consideration too.
Last edited:
And heed this advice too. Shoot for the datasheet current when you calculate R if you want to get close to the rated voltage.
Or something like this,
a true regulator with error correction. The addition of three components, the extra transistor and two feedback resistors create a proper regulator. The zener becomes a 12 volt one. The two resistors and caps to the left reduce the ripple component. The bits to the right are just for testing.
You can see how much ripple I've added to the unregulated input and how well the circuit reduces this under load.
a true regulator with error correction. The addition of three components, the extra transistor and two feedback resistors create a proper regulator. The zener becomes a 12 volt one. The two resistors and caps to the left reduce the ripple component. The bits to the right are just for testing.
You can see how much ripple I've added to the unregulated input and how well the circuit reduces this under load.
Attachments
As Dug said, you need a transistor rated for the full supply voltage. Also, because when you switch off, the collector voltage may fall faster than the base voltage, thereby reverse biasing the base-collector.
You need to connect a protection diode between base and collector (e.g. 1N4007). Another diode between emitter and collector is also good practice (some transistors already have this built in).
Last edited:
Mainly due to bad experiences with them..
I would argue about that "rugged" part, lol. Blew a few from fiddling and went over the max Vgs threshold. Should've put that gate-source protection zener..
Looking at my sand stash, i have IRFZ44N. Good thing is that it's TO220 and with G-D-S pin means it's easier to put in the TIP122 place (B-C-E pins). I'm just worried about the Vdss rating of 55 volt. I can imagine easily going over this at start up: Vd = 180v, Vg = 0v (waiting for the capacitor to charge) which means Vs also 0V.. so Vds = 180v. Clearly not OK? Or is it? I guess i have to get at least IRFP250?
Thanks for the circuit suggestion. I'm not sure i follow how it works, especially the values of the base voltage divider. The 2N5550 will start to conduct at 12v + Vbe = 12.7v and this combined with voltage divider of 14.7K and 120K mean it takes about 116v at the mosfet source for 2N5550 to conduct and limit Vgate? Quite far from 150V.. Or am i doing it wrong?
Also, i guess i can get away with skipping R2 and C2? The PCB is already crammed as it is. Adding another bulky high voltage capacitor is not what i want.
Last edited:
You need a much higher rated device than a 55v part.
The regulator is a classic "text book" implementation. Your description is spot on
its me that somehow had it in mind it was 120 volts output you wanted. You just alter the divider ratio to get the output voltage you want or replace R4 with a 22k preset and set it exactly.
The regulator is a classic "text book" implementation. Your description is spot on
its me that somehow had it in mind it was 120 volts output you wanted. You just alter the divider ratio to get the output voltage you want or replace R4 with a 22k preset and set it exactly.Might as well ask about the zener. Am i really that way off for the zener current?
180v unregulated - 156 zener voltage = 24 volt over the limiting resistor.
24 volt divided by 2K2 resistor gives about 11mA.
The 120V zener is ZY 120, a 2watt type which means i should aim for 2/120 = 16mA.
The 36V zener is C36PH, a 1 watt type giving Iz-max = 1/36 = 27mA.
Am i that far off? Do i really need to aim close to the maximum zener power dissipation to get good regulation?
Oh by the way, i haven't tested the circuit using full load. I use 100K resistor for the load. So it's only about 1.5mA of load current.. but i don't think this would cause the zener regulation to overshoot?
I would too, if i can find 2N5550 tomorrow.
180v unregulated - 156 zener voltage = 24 volt over the limiting resistor.
24 volt divided by 2K2 resistor gives about 11mA.
The 120V zener is ZY 120, a 2watt type which means i should aim for 2/120 = 16mA.
The 36V zener is C36PH, a 1 watt type giving Iz-max = 1/36 = 27mA.
Am i that far off? Do i really need to aim close to the maximum zener power dissipation to get good regulation?
Oh by the way, i haven't tested the circuit using full load. I use 100K resistor for the load. So it's only about 1.5mA of load current.. but i don't think this would cause the zener regulation to overshoot?
If I were you, this is the schematic I'd go for
I would too, if i can find 2N5550 tomorrow.
Listen
That circuit is a basic workable example. The FET type is fine, the transistor just happens to be in the LTspice library. You can use any general purpose type as long as the ratings are OK, particularly Vce. A 2n5551 is a common device, and an MJE340 should be fine too.
The zener isn't actually critical at all because you can just alter the divider ratio to suit. So 8.2volt zener, 12 volt, 30 volt even. All can be made to work. Four or five milliamps is usually plenty for a typical small zener.
Listen
I believe you from the start
I just want to know why my 120V + 36V series zeners give such bad voltage regulation so i know what NOT to do if i ever need to do something similar. Are zeners really that sensitive about the current?
Hi paul.. it's not that i don't believe you.. but what you're saying contradicts what i read here.
Here's from the article
It's actually the 120V zener that is quite far. I measured it gave approx 130v.. while the supposedly 36V zener gave 38-39 volt for a total of 168-169 volt.
I seem to recall from the dim and distant past, that for best stability diodes that were around 3.9 to 4.7 volts were preferred this being the critical voltage point at which zener diodes became more correctly "avalanche" diodes. We use the term zener for all of them but its not strictly speaking correct. Devices have improved over the years though.
A problem with your original circuit is that there is no error correction, its just a simple series pass regulator that sees the output voltage vary a little under different load conditions... which isn't necessarily a problem at all.
Why not test some zeners sorry, avalanche diodes yourself. Rig up a DVM up and try them from a cold start, testing with different currents up to their rating and write down the cold and hot voltage. Try it with several different voltage devices.
A problem with your original circuit is that there is no error correction, its just a simple series pass regulator that sees the output voltage vary a little under different load conditions... which isn't necessarily a problem at all.
Why not test some zeners
sorry, avalanche diodes yourself. Rig up a DVM up and try them from a cold start, testing with different currents up to their rating and write down the cold and hot voltage. Try it with several different voltage devices.No it doesn't. Your article just describes how to calculate a minimum resistor value based on maximum zener power when the load is removed. There's a maximum resistor value too, based on ensuring there is enough current through the zener for it to regulate. A good start for "enough current" is to use the device's test current.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.