• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Ribbon Microphone Preamp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I don't understand. You can't parallel two voltage supplies unless they are absolutely identical in all respects.
Why? What is wrong with following schematic?

If buzz occurs as soon as the DC heater supply has a load attached then you have coupling between the heater supply and the main circuit. This could be via wiring or via the mains transformer. You could try a choke input heater supply, as this will reduce charging pulses. The choke may be big, though!

There is coupling of course - heaters are elevated +90V from B+ via voltage divider 200K -- ( 100K || 220uF). BTW before I added that 220uF cap there was lots of buzz. But adding more capacitance here does not reduce it anymore.

Yea, chokes are big and expensive. I am trying to avoid it, if possible. Also I am not sure if it will help for sure. I do not have one to try.
May I replace it with using some (another) power transformer primary?
 

Attachments

  • parallel_PSU.png
    parallel_PSU.png
    21.3 KB · Views: 81
mm7 said:
What is wrong with following schematic?
If that circuit is what you have built, then one of the secondaries does no work as the other secondary will always present a higher voltage and so reverse bias the diodes from the first secondary.

You may have actually built it with R6 connected to the other end of R2 (top of C5). In this case R2 and R6 acts as current steering resistors or ballasts so both secondaries may contribute provided their voltage difference is not too great.

You may be injecting buzz into a ground. DC PSUs should always be grounded (i.e. connected to circuit ground) at their output end, not partway through as your diagram shows. That is, ground should be at C2, not C9.
 
That ground at C9 is just for LTSpice.
Actually in real heaters circuit there is no ground anywhere. It is somewhat floating being elevated.
Though it is couples with ground indirectly via elevation chain (voltage divider), via lower R9 and C16 of the voltage divider. Look at following schematic that reflects current state.

So what about using a power transformer primary as just a choke?
(in local Sayal they do not have chokes, but have plenty of trannies)
 

Attachments

  • PSU.png
    PSU.png
    33.6 KB · Views: 89
OK. If you apply a DC load to the heater supply, but with the heater supply disconnected from the potential divider do you still get buzz? You need to find what the coupling mechanism is: common ground currents, wiring magnetic induction, transformer interwinding capacitance etc.

A power transformer will make a poor choke because it cannot handle DC. You could rebuild it with an air gap.
 
...
So what about using a power transformer primary as just a choke?
(in local Sayal they do not have chokes, but have plenty of trannies)

If you get a big enough power transformer it will work well as a choke. The problem is a 100ma PT is rated for 100ma at 60Hz. If you want to put 30Hz through it the core size doubles, at 15Hz it is 4x larger. It is not infinite at DC but it's still big. All power transformers are designed such that their core is just barely large enough not to saturate at 60Hz and full spec'd current. Some might have larger safety factors than others. So find one with maybe 5X the mass of your PT? You can buy a real choke that is built with a gapped core from Triode Electronics. They ship to Canada which is only a few miles north of them.
 
... DC PSUs should always be grounded (i.e. connected to circuit ground) at their output end, not partway through as your diagram shows. That is, ground should be at C2, not C9.

This is right. Two comments.

1) is the ground is just to make Spice happy then place a 100M or even 1G resister in series with the ground. That will tell everyone to ignore the ground.

2) As said above the physical layout really does matter and I suspect some of the noise may be a result of things we just can NOT see in a schematic diagram. In fact coupling almost by definition takes place via paths that are not shown in the schematic.

In any case something is wrong here. I've built many amps that use AC heaters and the hum is acceptable. People usually result to using DC heaters only to remove the last tiny bit of hum. and when they do they usually find they need to regulate the heaters.

The neat thing about using an LM7806 for the heatrs is that the chip rejects noise better than any passive filter. (yes 6VDC is fine, maybe even better then 6.3) If you worry about cost, size and weight, adding one three terminal chip is better and more effective than a choke. But you have to feed it about 9V and heat sink it.
 
Last edited:
DF96>OK. If you apply a DC load to the heater supply, but with the heater supply disconnected from the potential divider do you still get buzz?

The elevator was connected to heaters.
When I've tested with battery heating, the elevator was connected through heaters to battery. No touch with heater supply.

When I've test with PS heating, the elevator was connected through heaters to PS. If I disconnect it, buzz gets loud.

DF96>You need to find what the coupling mechanism is: common ground currents, wiring magnetic induction, transformer interwinding capacitance etc.

Is there a way to find it?


Chris, I use star grounding with center at input XLR.
I have regulated PSU and I tried it. It buzzes even louder because it does not have a choke or even a resistor on input. High current into its input 15000uF cap overloads the transformer and creates buzz.
Four stages of RC filter smooths ripple to 0.3mV pp (in simulation).
Ripple does not come to B+ through elevator, it is directed to ground by elevator's 220uF cap.
It does not look as this ripple comes to cathodes from heaters. I added
one more stage with 15000uF. No change in buzz.
But what affects buzz is a load in heater circuit. If there is no load - no buzz. If there is load - buzz, even when heaters feed from a battery.
I've measured there is 300mA current between rectifier and first resistor. It is less than specified 1.5A.
 
Is there a way to find it?
....
But what affects buzz is a load in heater circuit. If there is no load - no buzz. If there is load - buzz, even when heaters feed from a battery.
I've measured there is 300mA current between rectifier and first resistor. It is less than specified 1.5A.

This is good information. It seems that if the diodes in the heater DC supply are switching then you get the buzz. Even if the supply is not supplying the heaters.

The only way I can think the heater supply rectifier can be causing this is if the coupling is via the power transformer. The switching noise is coupling either magnetically or capastivily through the PT.

I heard of people having trouble with DC heaters. The DC current is AVERAGE only 300ma but the diodes conduct for only (lets just say 1/100th of the time because your resivour cap is so huge) This means that while the average is 300ma the current goes for zero to 30 amps then back to zero. Yes is might be that large, you need an o'scope to see the conduction angle.


Had you tried an AC heater the current would be drawn from the transformer in much smoother way, certainly not as huge 30A spikes.

What to do? I'm guessing here but a smaller resfilter cap, perhaps 10uf or 22uf would cause the diodes to conduct over much of the AC cycle and stretch out the current spikes. Filter the DC with a second RC stage. I bet you'd get the same effect with a larger size (1K ??) resister in series with the transformer's heater secondary. Anything so the diodes will stay "on" for most of the AC cycle.

Or what the heck, just try using 6.3V AC heaters and see what happens. Yes, elevate it up a few volts with the voltage dividers. Just mack sure the rectifier is disconnected. If going AC helps you have found the problem.

One very good solution is to get a separate heater transformer then there is no way for the heater an B+ supplies to couple.

I have myself only had to solve this problem in the other direction. My B+ supply would conduct only over about 10% of the AC cycle and then of course the current drawn for the AC mains was pulsed, not continuous. So the current in the primary was not constant. The transformer has a 6.3 volt tap I used for heaters. Looking with a scope I could see the sine wive on the tube heaters has a HUGE amount of distortion. I think heater elevation fixed the slight buz.
 
The DC current is AVERAGE only 300ma but the diodes conduct for only (lets just say 1/100th of the time because your resivour cap is so huge) This means that while the average is 300ma the current goes for zero to 30 amps then back to zero. Yes is might be that large, you need an o'scope to see the conduction angle.
That is why I put 8.4R before the first capacitor. This reduces current to 0.5A pikes. And the more resistance I put the less buzz is. But I cannot put here to high R because voltage will drop too much.

Had you tried an AC heater the current would be drawn from the transformer in much smoother way, certainly not as huge 30A spikes.
No. Because I believe it will hum cathodes of upper tubes that are 40 volts higher than heater.

One very good solution is to get a separate heater transformer then there is no way for the heater an B+ supplies to couple.
Yeas. that will be my last resort.

Today I tried with feeding heaters circuit from 18VAC secondary (it can hold up to 1A). This allowed me to increase the input R to 15R, and other resistors in RCs to 3.3R.
The buzz decreased to -65dB.
Also I tried with my regulator where I put 9.4R before its input. Same, buzz decreased to -63dB.

Yeah. Next step will be a separate PSU for heaters.
 
That is why I put 8.4R before the first capacitor. This reduces current to 0.5A pikes. And the more resistance I put the less buzz is. But I cannot put here to high R because voltage will drop too much.
...


It looks like the cause is well understood now, coupling of DC supply back to B+

It would be educational to try elevated AC heaters and measure the buzz with AC.
 
It would be educational to try elevated AC heaters and measure the buzz with AC.

I tried. It is amazing! Buzz is less audible. But it is more hum.
May be due to sloppy wiring because DC was a target.
The heater wires are twisted, but there are inch or two untwisted wire between the sockets. Heaters between sockets are connected sequentially, with parallel connection inside a socket. So total Vh is 12.6V with 6.3 in each envelope.

Here is AC buzz/hum spectrum.
60Hz is louder, but higher harmonics are quieter, which sounds better for ears.
 

Attachments

  • Buzz_AC_10192013-06:25:36.png
    Buzz_AC_10192013-06:25:36.png
    38.6 KB · Views: 122
R8 and R9 in the PSU schenatic are probably too large and this may be contributing to the buzz. Most tubes like to have a heater cathode resistance of about 20K which implies reducing your values to about one fifth of their prenet values and increasing their rating appropiately.

In practice, with large amounts of smoothing capacitance,you do not get current spikes 100 times the average current. They tend to be limited by the dc resistance of the transformer secondary and the ESR of the capacitors. I have never measured spikes greater than 10 times the average current. Buzz caused by diode current spikes can be cured by bypassing them with 100nF.

Cheers

Ian
 
..
Here is AC buzz/hum spectrum.
60Hz is louder, but higher harmonics are quieter, which sounds better for ears.

I think now it is absolutely certain what the cause of this buzz is.

That is what I was getting at. Except in extreme cases AC can work fine. You could install a hum balance pot and better layout can maybe reduce the 60Hz by 3 to 9db. For most applications that is good enough. But for recording you need even better. That is when people resort to DC heaters. It is not needed in most normal HiFi amplifiers.
 
I targeted this preamp as a recording, not Hi-Fi.:cool: I know it is very naive for a first project. And it is ambitious and challenging, that is what I like for this hobby. So terms and costs are second priorities for this one.

What I've got from the AC experiment that even a large "ripple", which the 12V AC is, does not cause very loud hum when applied to upper tube's heaters that are 40V negatively "elevated". So DC with some mV ripple should be fine.

What creates buzz is some internal inductive coupling in the transformer. Or EMI from (over)loaded transformer? Not too much sure. I know if it is unloaded it does not buzz.

I am going to feed heaters from a separate power transformer. like one of these ones: Hammond Mfg. - Class 2 - Energy Limiting Transformers - ("BA" - "BE" Series)
though I do not quite understand what "energy limiting" means.
I like that they are boxed in shielding and are "[SIZE=+1]Rugged, low noise, cool running design".
I hope this will not produce more buzz (EMI) than the current one.

What do you think?

[/SIZE]
 
I think(?) power limiting means they are safe to short. Their internal resistance is so high that they can't burn up. I've not used this exact model but I've seen transformers like this used for door bells and lawn sprinkler timers. They are designed to mount to a standard steel electrical junction box. The AC main side mounts using standard conduit threads.

The version I've seen sell at the local hardware store and are rated very low, like 50 ma. These Hammond parts have bigger spec'd current. I've never seen a "bell transformer" as they are called used for audio

I have used "wall wort" plug in AC power supplies and then led the power to my project using a coaxial plug. The plug-in boxes are kind of the same thing. In fact my Art "Tube MP" preamp uses a 12VAC wall wort to power a 12AX7 tube.
ART Tube MP Studio Mic Preamp | Musician's Friend

Yes, who would have guessed that a 12 volt "ripple" would not sound so bad? I think your logic is right, cutting 12V to 12mV by going with DC should reduce the 60Hz hum by 60dB. The only thing working aginst that logic isthat rectified AC has a 120Hz, not a 60Hz fundamental. The 120Hz is easier to hear.
 
I have used "wall wort" plug in AC power supplies and then led the power to my project using a coaxial plug. The plug-in boxes are kind of the same thing. In fact my Art "Tube MP" preamp uses a 12VAC wall wort to power a 12AX7 tube.
ART Tube MP Studio Mic Preamp | Musician's Friend

Tried with wall wart 12V. Buzz/hum is not audible.
On spectrogram it does not go higher then hiss.

I think it is the way to go.
The only concern, for safety, is there a way to stop +90V elevating voltage to go to wall wart cable?
 
I plugged output into sound card E-MU 0404. And analyzed captured signal with Jack Audio Analyzer.
If you take a look at above post #351 you will see screenshot of the tool.
Currently I do not have buzz spikes. Hiss is on same place.
If I speak loudly or yell signal goes up to -20dB.
 
I plugged output into sound card E-MU 0404. And analyzed captured signal with Jack Audio Analyzer.
If you take a look at above post #351 you will see screenshot of the tool.
Currently I do not have buzz spikes. Hiss is on same place.
If I speak loudly or yell signal goes up to -20dB.

OK, that signal is the noise spectrum. It is not easy to work out the noise over the entire bandwidth from that spectrum. Just about all it is safe to say is that the noise level is not -90 to -100dB. If and its a big if, that spectrum show the noise per root Hertz, then to find the noise level over the entire bandwidth (such as you might read on a meter connected to the output) you need to multiply by the square root of the bandwidth. The square root of 20,000 is 141 which, converted to dB is 43dB. So, given the above assumption about the spectrum, then if the noise bumps along at -100dBu then the actaul noise over the whole 20KHz bandwidth is 43dB higher than this i.e. -57dBu

Cheers

Ian
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.