Don't know, never heard such noise with this reg. Change your cabling to the load, making it shortest, thickest possible and twisted, maybe its some oscillation or interaction. What do you power with it? What Voltage and current setting you use on the reg? Original transistor part numbers?
Also the pot can be dodgy as A.T. said.
Also the pot can be dodgy as A.T. said.
Audiohifi said:
See about tubes, see about grid stoppers, see about adding a 0.1uF COG ceramic or fast small PP like Wima where the wire comes to the preamp from B+ to ground, see about different wiring (thicker, different type & geometry, shorter). Do you have an Oscope? Look where this occurs, else we are stabbing in the dark.
Thanks to all for the pointers
The noise is heard when I have my ears close to my 86db speakers, almost no huming at all, only the transistor noise
I do not have a scope but can easily borrow one
How do I go about locating the source of noise? Will be glad if members can briefly outline the steps
Audiohifi
The noise is heard when I have my ears close to my 86db speakers, almost no huming at all, only the transistor noise
I do not have a scope but can easily borrow one
How do I go about locating the source of noise? Will be glad if members can briefly outline the steps
Audiohifi
Is it something like a ''hsh, hsh'' sound? Tubes do that sometimes too if there is intermittent oscillation or heater mechanical instability.
First exchange tubes. If you don't have new to try, exchange them between compatible positions-channels. Then, change the cabling from reg to pre as explained above. Add an RFI filter cap at the preamp receiving end as explained above. Up the grid stopper values or add some if there aren't any. If nothing of this works, get the Oscope and starting from the PSU line, going further to each plate and grid, check the display for noise transient each time you hear the sound. Have to sweep the sensitivity and time base until its visible.
Where it occurs is the source of the problem.
If you feel like checking out the reg first, add 100R base stopper on Q2. Maybe in your layout it needs it. Also check MPSA94.
First exchange tubes. If you don't have new to try, exchange them between compatible positions-channels. Then, change the cabling from reg to pre as explained above. Add an RFI filter cap at the preamp receiving end as explained above. Up the grid stopper values or add some if there aren't any. If nothing of this works, get the Oscope and starting from the PSU line, going further to each plate and grid, check the display for noise transient each time you hear the sound. Have to sweep the sensitivity and time base until its visible.
Where it occurs is the source of the problem.
If you feel like checking out the reg first, add 100R base stopper on Q2. Maybe in your layout it needs it. Also check MPSA94.
Yes, what salas said.
First thing I'd do, is make the connection between regulator and amp very short. Those long wires are great antennas. In fact when I built my first tube headphone amp I could very nicely hear a local radio station and enough shhhhhhh type of noise. Your regulator is pretty much defeated with long wires. A small cap (100nF ceramic) right at the point of B+ on the amp also worked for me to reduce some of the RF noise. But best is to have short cables.
If you get the scope, you will have to be very careful with the probes, you're working with very high voltage. I would suggest you do not unhook/re-hook the probe while the circuit is powered. You can get hurt and/or fireworks with a slipped probe. This being said, you would check for noise (set the scope on AC and appropriate time/div and V/div settings, probably 100mV/div first) as salas said. But note that wherever you do find the noise, it may actually be generated a stage before where you find it, but be too small for your scope to see. By the time it gets to your speakers it's been amplified probably a few times.
My 2c.
First thing I'd do, is make the connection between regulator and amp very short. Those long wires are great antennas. In fact when I built my first tube headphone amp I could very nicely hear a local radio station and enough shhhhhhh type of noise. Your regulator is pretty much defeated with long wires. A small cap (100nF ceramic) right at the point of B+ on the amp also worked for me to reduce some of the RF noise. But best is to have short cables.
If you get the scope, you will have to be very careful with the probes, you're working with very high voltage. I would suggest you do not unhook/re-hook the probe while the circuit is powered. You can get hurt and/or fireworks with a slipped probe. This being said, you would check for noise (set the scope on AC and appropriate time/div and V/div settings, probably 100mV/div first) as salas said. But note that wherever you do find the noise, it may actually be generated a stage before where you find it, but be too small for your scope to see. By the time it gets to your speakers it's been amplified probably a few times.
My 2c.
I have only recently started reading this thread, and I fancy trying this regulator with my valve active crossover, which needs 300V at around 80mA.
I have been trying to work out how the circuit works, and there is one aspect that I don't quite understand. The voltage across the base-emitter junction of the high-voltage PNP transistor Q2 is equal to the base-emitter voltage of Q4 plus the total voltage across Q5. In other words, either Vbe of Q2 is going to be a lot higher than 0.6V, which I would expect to damage the transistor, or there is almost no voltage across Q5, which can't be the case, since it is carrying something like 4mA. What is going on?
Alex
I have been trying to work out how the circuit works, and there is one aspect that I don't quite understand. The voltage across the base-emitter junction of the high-voltage PNP transistor Q2 is equal to the base-emitter voltage of Q4 plus the total voltage across Q5. In other words, either Vbe of Q2 is going to be a lot higher than 0.6V, which I would expect to damage the transistor, or there is almost no voltage across Q5, which can't be the case, since it is carrying something like 4mA. What is going on?
Alex
AndrewT said:
It seems Allen Wright used to believe that too, but says he has changed his mind recently - he now recommends about 20% of load current with his SuperReg. It also, of course, depends on the stability of the AC mains voltage.
If the shunt current is 20% of the load, and the input voltage excess is 20% of the regulated output voltage, then the shunt MOSFET will dissipate 20% of the dissipation in the load, or 4.8W, which is manageable, as is the 5.8W in the CCS MOSFET.
Alex
Hi Andrew, Allex,
I wonder if a 20% of load thru the shunt would work for my F3 amp - it consumed about 1.7amps at 42 volts, so would end up with about 2amps thru the reg with a CCS of about 10W and the Shunt about 14W (42 x 20% of 1.7) for a total of 24W.
I like the idea but a shunt consumption of at least 1.7 amps@42volts is quite impractical, so put it aside.
I wonder if a 20% of load thru the shunt would work for my F3 amp - it consumed about 1.7amps at 42 volts, so would end up with about 2amps thru the reg with a CCS of about 10W and the Shunt about 14W (42 x 20% of 1.7) for a total of 24W.
I like the idea but a shunt consumption of at least 1.7 amps@42volts is quite impractical, so put it aside.
You must allow for peak + 20%. Is your amp using 1.7A at peak when the oscope shows the onset of flattening on a max power sinewave?
Then yes, you can use a shunt at 2A and be OK, with 0.3AX42V constant excess on the second Mosfet. Class A amps are the best customers for constant current shunts, since they have the big sinks and burn the most juice there for bias all the time. Just don't try a test without 1.7A load with the 22.6W CCS & Shunt dissipation in mind.
Then yes, you can use a shunt at 2A and be OK, with 0.3AX42V constant excess on the second Mosfet. Class A amps are the best customers for constant current shunts, since they have the big sinks and burn the most juice there for bias all the time. Just don't try a test without 1.7A load with the 22.6W CCS & Shunt dissipation in mind.
Thanks Salas,
Got a "bee in the bonnet" about having more current thru the shunt than load - I remember we talked about it awhile ago.
However, I just might bang a bunch of bits together on some of that perf board and crank it up and see - I think the F3 will be transformed with it, if it works okay and the heatsinks should accept another 20+ watts (the standard 80 Watts = 22*C rise).
I have done just the one pcb and it's checked out A-okay, no problems, except still a bit too small for me (will do a much bigger design, next time) - don't know why I added all those options, as few enquiries - I had hoped someone over your side of the globe would copy it into "Eagle" and make it more user friendly. I haven't got the hang of that one at all yet.
I might just try out Iko's new V1.5 with just standard bc transistors and see what happens on the DAO headamp supplies - the current V1 with it's heavy shunt current (250mA load, 300ma shunt) is working flawlessly - no fatigue at all, yet incredible dynamics and detail plus effortless bass control with a clearly defined sense of rythmn - my sort of good sound, and not such any easy thing to manage on any equiptment.
Got a "bee in the bonnet" about having more current thru the shunt than load - I remember we talked about it awhile ago.
However, I just might bang a bunch of bits together on some of that perf board and crank it up and see - I think the F3 will be transformed with it, if it works okay and the heatsinks should accept another 20+ watts (the standard 80 Watts = 22*C rise).
I have done just the one pcb and it's checked out A-okay, no problems, except still a bit too small for me (will do a much bigger design, next time) - don't know why I added all those options, as few enquiries - I had hoped someone over your side of the globe would copy it into "Eagle" and make it more user friendly. I haven't got the hang of that one at all yet.
I might just try out Iko's new V1.5 with just standard bc transistors and see what happens on the DAO headamp supplies - the current V1 with it's heavy shunt current (250mA load, 300ma shunt) is working flawlessly - no fatigue at all, yet incredible dynamics and detail plus effortless bass control with a clearly defined sense of rythmn - my sort of good sound, and not such any easy thing to manage on any equiptment.