|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.13.28.146
Has anyone replaced the half-wave C- power supply in any of the many vintage tubed amps that use them with a full-wave supply? What were the results?
We're doing a rebuild on a McIntosh MC240 and are strongly considering it. It's cheap and easy, and it seems to me it can't be worse than a half-wave supply.
Follow Ups:
C- supplies are typically half-wave because the extremely low current draw doesn't benefit from full-wave rectification. Even with ripple at the line frequency (as opposed to twice that frequency with full-wave), capacitor filtering is more than adequate. So, on that front at least, there wouldn't seem to be a reason to change the topology.
On the subject of regulation (mentioned elsewhere in this thread), it's possible to add an active circuit that "cleans up" the bias voltage, but doesn't regulate it. This approach would provide the dynamic benefit of a regulator, but without the disadvantage of skewing operating conditions with changes in line voltage. Below is a simple circuit of this type. Just something to think about.
--------------------------
Buy Chinese. Bury freedom.
If you put a scope on the negative bias supply rail while music is playing,you can see the audio there at the source riding on the filter caps.
Simple logic would tell us that some of that audio signal is probably getting filtered out in the C minus supply since you can see audio riding on the C minus rail. Even tho bias draws very little current and needs very little filtering,wouldn't it make sense to build the lowest impedance C minus rail that you can so you have less of a chance of this happening? When I put a toroid and a full wave bridge in a Citation 2 in place of the original one diode C minus rail,it is noticeably better IMHO and I was almost tempted to add a choke in line as well.Maybe I'm over analyzing this but some of the best home brew and commercial amps I've ever heard use battery bias in the negative bias supply. The last one I heard used series connected nicads..
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
"If you put a scope on the negative bias supply rail while music is playing,you can see the audio there at the source riding on the filter caps. [snip] ...wouldn't it make sense to build the lowest impedance C minus rail that you can"I really want to answer in the affirmative, because this seems so intuitive. However, I think the question indicates a possible misconception about the cause and effect of the observation.
To analyze the cause, it's necessary to first consider the relationship between the C- filter cap and the audio signal at the power tube grid. Regardless of how that signal arrives at the grid, it's decoupled from the bias circuit through either a high value resistor (150K, 220K, etc.) or a high value of inductive reactance (choke or transformer). Both methods serve the same purpose - to isolate the signal on the grid from the impedance of the bias supply.
In the case of resistive decoupling, the effect can be reasonably well quantified. For example, let's say the bias supply filter includes a 47uF capacitor to ground, and the cap is connected to a 220K series resistor that delivers C- to the grid. If this RC is analyzed by applying a 100V P-P signal at 20 Hz (worst case frequency) to the resistor, only 56mV P-P appears across the cap. That's a reduction of more than 62dB, and that number doesn't include the additional attenuation created by any of the other bias supply components.
As a result of this analysis, I think it's very unlikely that grid signal comprises the majority of what you see on the scope. If it is, the simplest remedy would be additional C- capacitance. However, there's a subtler path for audio to the top of the C- filter, and it's more likely to be the cause of the effect. Specifically, when most power stages are modulated by signal, they in turn modulate the B+ supply. This creates voltage and current fluctuations that reflect back through the supply and into the power transformer. The transformer, in turn, modulates the C- supply voltage. This represents a source of audio that acts on the C- supply at a much lower impedance than the power tube grid, and for which additional filtering across the supply will likely be less effective.
So, the short answer to your question is that a C- supply exhibiting a lower impedance to ground will indeed reduce audio signal appearing on the C- filter cap. However, reducing the impedance between the cap and the power transformer probably won't. The latter is what (marginally) occurs when the supply is changed from half- to full-wave rectification.
--------------------------
Buy Chinese. Bury freedom.
Edits: 07/10/16 07/10/16
I have done this on several Citation 2s and I like the sound very much but in the case of the Mc240 and other Mac amps,we have a different situation.Mcintosh uses a bias tracking system and it works in conjunction with the power transformer primary and secondary meaning,when the voltage goes up or down,the proportionality of the bias goes up or down to maintain the correct bias.You always set the bias on a Mac amp with a variac and a voltmeter.On an Mc240,you set the variac to run the B+ 440vdc on the plate and minus 47 on G1. If you don't have minus 47,you adjust the bias resistor up or down until you get minus 47 at 440v.This way when you plug it back into a normal AC wall voltage,the amp will track perfectly regardless of where the voltage moves.
What I do on Mc240s is exactly what McShane did on the Citation 1 and the Citation 5..You put in the 820uf doublers and than dual 100ufs down stream,along with a pair of dual 50uf paralled in for added decoupling capacitance..Then add a bussbar and bolt right on to the choke bolts..I also use Schottky diodes but the GI diodes are fine. The other thing I do is add a hash filter that Eli came up with and that gives you an even quieter background.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
Edits: 07/03/16 07/03/16
I don't understand. I'm not talking about regulating the C- supply, just using a bridge to rectify (and a resistive divider to establish bias). We're not going to regulate C- or B+, just filter the crap out of both.
I understand the inadvisability of regulating C- without regulating B+. Did you misunderstand our intent or do I misunderstand what you are saying?
Lee
I guess I misunderstood because when we do a full wave bridge on a Cit 2 bias,we end up using a torroid with dual secondaries that we series connect and then use a 4 diode bridge.
To be honest,doublers in the Mac and Cit 2 amps regulate very well by virtue of their low resistance secondary windings especially when you use it with the new doubler caps they have today.If you use the 820uf panasonic or Nichicon 105 degree caps along with a good buss bar and an added hash filter,you aren't going to do any better than that for filtering.I suppose there are other ways to do it,but Jim McShane has tried many different setups before he came up with the one for the Citations and we just implemented with the Macs because the power supplies mirrored each other. The added decoupling also helps with stability.I don't know if you have seen some of my posts but I have always put up the end results with measurements and waveforms and that kind of gives you an idea of where you are at..You can try a different way if you keep the same taps..I thought you were using a different bias transformer and with the Macs you can't if you want it to track correctly.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
Edits: 07/03/16
Ah, we are on the same page.
I don't think "regulate" is the word you're after in "...doublers in the Mac and Cit 2 amps regulate very well..." since the voltage will still scale with line voltage. I recognize the quality added by low resistance windings.
I was speaking hypothetically when I said regulate but what I really mean is stability..Today's Mac tube amps like the reissue Mc275 doesn't use any form of regulation that we use in conventional circuitry.It essentially mirrors the original circuit. The reason behind that I would assume is the fact that the unity coupled circuit doesn't operate independently of the other elements in the circuit.It moves proportionally up or down with signal. In a conventional pentode connected output circuit, we can regulate G2 without regulating the plate voltage.
I am curious tho. I would love to see what would happen in a Mac amp if we did regulate the plate and or G2 just to see what would happen to operating conditions.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
I think you'd be ok if you regulated B+ AND C-, but not just one. Since G2 and the plates are connected, if you regulate one of them I think you'd drastic change the operation of the amp.
I believe you are right.I may give it a try as well.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
generally better sound.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: