I have an EAR834p clone, and the power supply died (burnt resistors, failed zener). I am rebuilding it. I am good with the standard PSU filtering, but this is my first time working with a silicon voltage regulator Transistor and a zener stack.
I am getting a situation where the regulated voltage ramps up to 285 pretty fast, but then continues to creep up past 310V. Zener stack is hot too. With 72V each across the 68v rated zeners. I suspect that is what led to failure.
I am getting about 360V b+ unloaded going to the Transistor, an FJP13009.
The FJP13009 data sheet says a minimum of 400v. Does that mean my input B+ is actually too low to be properly regulated to 385?
On the choke will be approximately 355v,
on the collector of the regulator also.
There is (according to your schematic) 285v on its base, meaning there should be 284 ish on its emitter as a regulated supply due to the zener chain on its base.
That means the ripple on the collector can be as much as 72v before ripple will arrive on the emitter and the voltage drop across the resistor network (3k + 3k = 6k) means a current of about 12mA feeding the zeners.
12mA is not enough to force overvoltage across the zeners alone what are they and why have you got the wrong voltage across them ...
I suggest the transistor has a Collector Base leak.
It cannot regulate at 385 because there is only 353v available from the 250v AC supply; 250 X 1.414 = 353.5 hence my 350v approximately notation.
Just because a transistor data sheet states a maximum 400v Ce doesn't mean it must have that available to work as described.
on the collector of the regulator also.
There is (according to your schematic) 285v on its base, meaning there should be 284 ish on its emitter as a regulated supply due to the zener chain on its base.
That means the ripple on the collector can be as much as 72v before ripple will arrive on the emitter and the voltage drop across the resistor network (3k + 3k = 6k) means a current of about 12mA feeding the zeners.
12mA is not enough to force overvoltage across the zeners alone what are they and why have you got the wrong voltage across them ...
I suggest the transistor has a Collector Base leak.
It cannot regulate at 385 because there is only 353v available from the 250v AC supply; 250 X 1.414 = 353.5 hence my 350v approximately notation.
Just because a transistor data sheet states a maximum 400v Ce doesn't mean it must have that available to work as described.
Sorry, but are you suggesting that there is an issue on my board causing a leak (like a faulty solder joint?) or a problem with the Transistor? I just replaced the Transistor (used to be a MJE13009 when it originally failed, now replaced with a FJP13009 to rebuild it).I suggest the transistor has a Collector Base leak.
It cannot regulate at 385 because there is only 353v available from the 250v AC supply; 250 X 1.414 = 353.5 hence my 350v approximately notation.
Or is it that I need at least 385V on the choke to get to regulation?
John is suggesting that there is too much current flowing through the zener string, and is caused by current ‘leaking’ (improperly flowing) from the pass transistor’s collector to its base. Indicating, the transistor needs replacement, as there should no such current flowing. Even with a good transistor, the zener string is conducting more current (and therefore, dissipating more power) than is necessary. Meaning, the 3k and 3.3k resisters could be increased some. Higher than needed power dissipation necessarily reduces the reliability of a device.
Last edited:
Thank you. I will try that. I admit I dont really understand what those two resistors do...is it just a voltage divider to get 285v from the B+ on the transistor base?Meaning, the 3k and 3.3k resisters could be increased some. Higher than needed power dissipation necessarily reduces the reliability of a device
My mains power is 120V, rather than the 115V for the transformer primary. So my B+ coming off the secondary is higher than the 250V speced. Perhaps that is causing the resistor voltage divider to set the Voltage at the base higher than intended? And this higher current draw on the zeners?
The two resistors, combined with the two 47uF capacitors, form a 2nd order low-pass filter. The resistors also proved the current path for biasing the zener string. A string of four 68V Zener should give 272V (68V * 4), and not be over-biased to produce 285V via too much current. Which, presently computes to (360V - 285V = 75V) / 6.3k = 12mA. More than is needed. The zener string shouldn’t require more than 5mA of bias current, likely less, to provide a voltage reference for the transistor’s base.
The zener string’s bias current is set via the 3K and 3.3K resistors. With 360V on the transistor’s collector, and a 272V target on the zener string, The two bias resistors combined value should be around (360V - 272V = 88V) / 5mA = 17.6K. Which is only a target. Spilt the chosen value in half to arrive at the the value for each of the two bias resistors. Using the example, (17.6K / 2 = 8.8K) for each.
Last point is the the max collector to emitter voltage of the MJE13009 is 400V. This is enough after the regulator has stabilized, but during initial turn on, the collector-emitter is briefly exposed to 360V as filter I had mentioned earlier charges up. keep in mind, that the charge up time will increase, as the bias resistor’s value is increased. This could potentially necessitate a reduction in the values of the two 47uF capacitors.
The zener string’s bias current is set via the 3K and 3.3K resistors. With 360V on the transistor’s collector, and a 272V target on the zener string, The two bias resistors combined value should be around (360V - 272V = 88V) / 5mA = 17.6K. Which is only a target. Spilt the chosen value in half to arrive at the the value for each of the two bias resistors. Using the example, (17.6K / 2 = 8.8K) for each.
Last point is the the max collector to emitter voltage of the MJE13009 is 400V. This is enough after the regulator has stabilized, but during initial turn on, the collector-emitter is briefly exposed to 360V as filter I had mentioned earlier charges up. keep in mind, that the charge up time will increase, as the bias resistor’s value is increased. This could potentially necessitate a reduction in the values of the two 47uF capacitors.
Thank you Ken for the math. I think I finally understand what is going on here. I will try adjusting those two resistors up to 8k and remeasure.Which, presently computes to (360V - 285V = 75V) / 6.3k = 12mA. More than is needed. The zener string shouldn’t require more than 5mA of bias
This is my first time working with zeners and a transistor regulator. I have used 03C and 0D3 tubes as shunt regulators in a psu many years ago with my Aikido preamp, and I assumed this circuit worked similar with a voltage needed to "fire" the Transistor to engage at an operating voltage, which then settled at the target voltage. But now I realize that doesn't make sense here.
jayme, any update after you installed the 8K resistors? I just finished building the same amp.
The zeners are running 180+ degrees F. Too high for my liking, but the HV is normal at 285.
The zeners are running 180+ degrees F. Too high for my liking, but the HV is normal at 285.
yes, the pair of 8k resistors worked well to reduce the current into the zener string. I am not sure what the temperatureof the zeners runs at tho. i have not tested.
i also replaced my torrodial transformer with an R-core one reducing from 250V secondaries to 220V. i found this was needed when I switched from the original CRC to CLC design. I used the Hammond 155J (15H at 30mA ADC).
my B+ is now solid at 285.
the weak link is now the Alps pot on the preamp board. Anything other than fully open significantly degrades the sound. It was a lot less noticable when the B+ was funning hot around 300V.
i think it is overall my clear and musical at 285V, tho.
i bought another board, so that when I finally stop tweaking, I can build a cleanly soldered version and be done.
i also replaced my torrodial transformer with an R-core one reducing from 250V secondaries to 220V. i found this was needed when I switched from the original CRC to CLC design. I used the Hammond 155J (15H at 30mA ADC).
my B+ is now solid at 285.
the weak link is now the Alps pot on the preamp board. Anything other than fully open significantly degrades the sound. It was a lot less noticable when the B+ was funning hot around 300V.
i think it is overall my clear and musical at 285V, tho.
i bought another board, so that when I finally stop tweaking, I can build a cleanly soldered version and be done.
Thanks jayme for the update. Will get the 8K installed, on next parts order. I also used the choke in place of the 1K resistor. Now both the choke and resistor are installed in series. This did help lower zener temps about 20 degrees. This was built with the supplied 260 volt rated toroidal, and it measures 300 volts AC going into the board.
Reading through several hundred posts about this amp, the alps pot does degrade the SQ. I chose to not install it on this build. Some builders have success installing a stepped attenuater. But at this point I'm not installing anything.
Reading through several hundred posts about this amp, the alps pot does degrade the SQ. I chose to not install it on this build. Some builders have success installing a stepped attenuater. But at this point I'm not installing anything.
- Home
- Amplifiers
- Power Supplies
- Voltage Regulator - question about Transistor+Zeners on EAR834p clone