It's one of those things that will bug me if I don't. I am working on an amplifier that uses the same +/-55v rails for the power amp section and preamp section. I was planning on giving the preamp its own power supply to cut noise down. To do this, I wanted to give the preamp its own capacitor bank and current source, unregulated of course. Right now the +/-55v is created from a center tapped transformer, center tap grounded. Is there any reason I can't connect two rectifiers to this one secondary to separate the power amplifier and preamplifier DC supply? I don't think it will work but I can't seem to find a sure answer anywhere else on the net. I need another transformer right?
If I do need another transformer to give the preamp its own rectifier, is there a simpler way of separating the two capacitor banks so the power amp can't draw from the preamp caps and vice-versa? Perhaps connect the preamp capacitor positive to the rectifier by diode and place the load after this diode? Or would this also be a bad idea?
If I do need another transformer to give the preamp its own rectifier, is there a simpler way of separating the two capacitor banks so the power amp can't draw from the preamp caps and vice-versa? Perhaps connect the preamp capacitor positive to the rectifier by diode and place the load after this diode? Or would this also be a bad idea?
A separate transformer is best, but if the amp and preamp are in the same enclosure, you might be able to work out the details to have low hum and noise. Most receivers use only one transformer, and they work well enough. The diode decoupling idea has been used by Audio Research in their tube power amps, between the output stage and driver stage, but they added lots of capacitance at the driver.
With two rectifiers on the same secondary, when the amp draws large current peaks, this will pull down the AC secondary voltage, and the preamp's rectifier will be cut off at those times. You'll need lots of capacitance after the preamp rectifier to hold up its output voltage well enough.
Last edited:
I figured that since rectifiers draw most of their current from the peaks of the AC waveform this would probably be bad practice even if it worked. So I take it take it the diode idea isn't unsafe? I may as well experiment with it. I will be ordering large caps for both supplies since they are unregulated and so, as I understand it, why not? Going from 8,200uf x2 shared per channel to 4,700uf x2 per preamp channel and 15,000uf x2 on the output. I should be more specific about the "preamp" board since it is the line stage (preamp, right?) and voltage driver.
Bear in mind that the diodes can only handle a certain amount of peak current, which will be increased by the increased
capacitance. The safest device to use would be a 30A or larger diode bridge rectifier. These have 300A peak rating.
Many higher power integrated amps have been built commercially and have only one transformer but nearly always have a separate winding for the preamp power. That's mainly because preamps run more tame power supply rails - split 12 or 15 because that is the common limits for the opamps.
To get the 55s down to 15 will need a special regulator to handle 40+ Volt drop and the heating will be a problem. Also, your 55 volt rails can exceed 65 depending on the power company. They don't do a great job of controlling line Voltage swings. Just things to keep in mind.
G²
I think this is a little more than a preamp, it is the output stage driver. I am not building this amp, it is a pioneer A-77X. I live in Florida and FPL provides us with 119.9v at 60.1hz bang on 24/7 unless the storms pop a transformer or something locally.
And the line Voltage doesn't drop when the neighbors A/C fires up? My lights flick a little when the next door A/C turns on. I get a much larger flick when my own A/C turns on. Of course things like this happen in southern California.
I'm pretty sure if you ran a chart recorder on your line Voltage there would be a little more spread throughout a 24 hr period. The point is not everyone has nice stable line Voltage and it must be accounted for in designs.
G²
I have a new AC with a new start capacitor so there is no fluctuation when the AC turns on, my neighbors AC doesn't seem to do anything either. FPL does fairly well, besides the reliability in bad weather, but that is any above-ground system. This amp will run for a few seconds off of its caps right now so I assume you would need some pretty major fluctuation to cause any problems.
Did anyone suggest an RCRCRC filter direct from the smoothing caps of the mains supply?
Select the R to drop voltage at the expected current draw.
Make the final C adequate for the peak current draw.
some suggest 2000uF per amp.
but that means only 100uF for a 50Apk of current.
I would put in 1mF (ten times as much)
The preceding caps could be sized at 1mF to 2mF per amp.
Select the R to drop voltage at the expected current draw.
Make the final C adequate for the peak current draw.
some suggest 2000uF per amp.
but that means only 100uF for a 50Apk of current.
I would put in 1mF (ten times as much)
The preceding caps could be sized at 1mF to 2mF per amp.
I believe a similar filter is already present on the board, made with some resistors and 2x 47uf 63v caps + 2x 100uf 63v caps per channel. I am thinking at this point that I will go with one rectifier per channel made up of Vishay fast diodes then use regular diodes to decouple each supply, four diodes total one for each pre and power amp channel. I will listen with various different stages decoupled and see what sounds best. Hopefully the benefits of isolating the supplies will outweigh the noise gained from the diodes.
Last edited:
House fire?
Rectifier (-) is -55v I would connect this the same way with diodes. The return to ground would of course be unimpeded with additional diodes, no need for that. I assume I will need to pick diodes with high enough current ratings but this should work, correct? If I experience some small forward voltage drop, it is tolerable as long as the transistors can switch on, correct? Going from 55v to 53v isn't going to pull too much current across anything? This supply is unregulated.
Rectifier (-) is -55v I would connect this the same way with diodes. The return to ground would of course be unimpeded with additional diodes, no need for that. I assume I will need to pick diodes with high enough current ratings but this should work, correct? If I experience some small forward voltage drop, it is tolerable as long as the transistors can switch on, correct? Going from 55v to 53v isn't going to pull too much current across anything? This supply is unregulated.
Last edited:
Before suggesting do this or do that I'd first like to know what your actual problem is, whether you actually have a problem and if so, what's its cause.
I find it very unusual that a preamp is fed +/-55V , doubly so if that's exactly the same power amp rail voltage, which makes me think that even if +/- 55V are sent to the preamp board, they are most certainly filtered, regulated and padded down to more sensible levels.
Which by the way destroys hum and ripple, rendering your extra rectification and raw filtering unnecessary.
What is your preamp made of?
If Op Amps, measure the actual rail voltage.
I find it will be around +/- 15V as stated earlier.
And if discrete, it might be higher, but definitely not same as power amp valuies; it would be a very poor design and can't believe an established Audio company would do that.
EDIT: now I read:
Well, then that is NOT the preamp but the power amp gain and drive stage.
90% of the POWER amp, by the way.
*IF* you want to separate it from main rails ripple and pumping, you do not need separate rectification, simply separate + and - rails with a couple diodes and add extra capacitance after them.
That way, when main rail voltage drops under load and ripple appears when it gets recharged, diodes become reverse biased and isolate drive stage power from that.
This is what I am seeking to do. I realize this is most of the power amp, the output stage being only a current amp as with most commercial designs. I was just wondering if NEGATIVE DC voltage "flows" through a diode anode to cathode like normal, or if I should reverse them.
If you are talking about putting a diode in the -55V leg of the DC supply, it would be reversed compared to one on the positive side.
The band end (cathode) of the diode would connect to the -55V, and the non-band end (anode) would go to the circuit being powered.
It's the current that flows through the diode, not the voltage, but this can get confusing. There is the conventionally defined (positive) current,
and also the actual (electron) current. They flow in opposite directions. Hole current in semiconductors can make all this even more confusing.
The polarity of the current is more or less an historical artifact, and a matter of convention.
Voltage is a quantity defined as the energy per electron charge (Joules per Coulomb) at a particular point in the circuit, compared to some other point,
which is usually common or ground. Voltage can also be measured relative to some other node in the circuit which is not connected to ground.
This type of measurement is called a floating measurement, and requires a probe that has neither side connected to ground.
Most DVMs can do this, but oscilloscopes usually require a special probe to do floating measurements, since their BNC inputs are grounded for safety.
Last edited:
Hence my putting "flows" in quotes in my original post. Thank you for the advice, I just wanted to play around with diode decoupling these supplies to see if there is a difference in sound.
You might be able to hear some difference on music with strong bass that would exercise the diode decouplers,
by dropping the main supply voltage by a few percent.
So far I cannot tell a difference with or without them on the pre/voltage card supply. I need some caps to put behind them, voltage has been steady at 54-56v (running at operating temp after 3hrs on) before the diodes, 53-55v after them. I haven't tried decoupling the output stage(s) yet.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.