Modding an Alctron Tube Mic PSU: What is EF86 B+ test load resistor?

[SparkleBear]

Your schematic makes no sense to me.

I was just following the technical documents to make LM317 output negative voltage (for u67 heater). This involved Grounding the output and the adjust pin now becomes the new output at a negative voltage... still using the same resistors to tune the output.



This follows what RuudNL changed about my schematic above too... except then i went overkill with the capacitors lol.

In my first incarnation of the psu mod i followed what you said with this:



But then learning i needed negative voltage i just moved the filters to the new output line (adj) and flipped the polarized caps.

Ok, so I will return to my original idea of 10ohm and 1000uf RCRC for the filter. ( I accidentally got 10,000uf at the electronics store )

Is there a better idea for getting negative voltage for the heater supply? Should I just go completely unregulated and just drop down from the 20v negative side of the diode bridge?

 

[rock soderstrom]

I was just following the technical documents to make LM317 output negative voltage (for u67 heater). This involved Grounding the output and the adjust pin now becomes the new output at a negative voltage... still using the same resistors to tune the output.

Your design has fundamental problems, you got something wrong. We all do that from time to time.

In essence, all was said (Ruud). You add to the existing circuit the additional filtering as you had it already apparently. But then your filament voltage was still positive, which is of course not good for the following circuit, which wants to generate the negative grid bias voltage from it. (almost a miracle that this worked at all).

To solve the problem, simply reverse the polarity of your heating PSU, i.e. the ground goes to the former positive output of your filament PSU and you pick up the heating voltage at the former ground, so it becomes negative 6,3V. Done!

Ruud was successful with this and I have no reason not to believe it.

Of course you can also generate the 6.3VDC with brute force method, it will work but I would first go the way mentioned, because that offers some advantages.

 

[Mr. Moscode]

What is the R value in your B+ RC string? I recommend some sort of regulation-Zener, active or zener, a mosfet and constant current sources. WHY? Because line voltage variations WILL show up in an unregulated B+ supply and will show up at your mic output. I have designed very high gain phono preamps with 80db of gain at 1khz there is a gain of 100 db at 20 hz due to the RIAA eq network. So regulation is important in my world, anyway.

To see this effect plug your PS with the mic attached into a variac, scope the output, and jiggle the variac setting up and down by a few volts. Add a zener a few caps down and repeat. The wiggle will be attenuated.

That low frequency bump may cause momentary saturation of the mic's output transformer. So what are the R and C values of the filter network? BTW, thats a very slow delay line.

MM

 

[SparkleBear]

Your design has fundamental problems, you got something wrong. We all do that from time to time.

Hehe, this is my first go at making my own tube mic PSU so I expected to make some mistakes. Thanks for the encouragement and support!

I scrapped that design and now have just gone with the simple swapping of ground and heater line, flipped around the filter caps and made a successful and stable negative heater voltage rail! The psu seems to be mostly in order. Just need to get a 220v diode and strap it between B+ and GND. The B+ with mic connected DOES ramp up to around 260v before it heats up and goes down to the nominal 210v. I thought that 220v would be the perfect setting.

Thanks so much for your help Rock and Ruud. (kind of a catchy band name.. wink wink)

What is the R value in your B+ RC string? I recommend some sort of regulation-Zener, active or zener, a mosfet and constant current sources. WHY? Because line voltage variations WILL show up in an unregulated B+ supply and will show up at your mic output.

Yeah, I am planning on getting the 220v zener in there to clamp the voltage and keep the capsule safe. Hopefully once it mellows out back to 210v it wont be effecting the circuit anymore... I have never designed one of these before so its all new and learning...

Here are the values:

Of course you can also generate the 6.3VDC with brute force method, it will work but I would first go the way mentioned, because that offers some advantages.

Can you share experience about any of the advantages and disadvantages of regulated vs unregulated heater supply? B+ supply? This is a topic of curiosity and have heard mixed opinions.

CHEERS!!!

EDIT: The voltages appeared to be fine but actually am experiencing a strong BUZZ when the PSU is turned on and connected to the microphone.. When the tube heats up and I turn the psu OFF it goes away and i can hear a clear signal for a moment as the tube cools down. Again, all the voltages look great but the hum is undeniable and unusable. I have retested the mic with my DNU67 and its definitely the PSU... HMMM... It sounded way better with the configuration with positive heater voltage lol... What about using an LM377, the negative version of the LM317? What to do....

 

[Mr. Moscode]

The advantage of a regulated power supply is stability and power line immunity. You can use 2 stage regulation. It will be very quiet without a lot of capacitors. In addition the voltage to the amplifier will be pure, low impedance DC no matter what frequency the amplifier is drawing keeping frequency response of the gain stages independent of power supply reactive elements, i.e. capacitors. Also, keeping the heater voltage stable keeps the operation of the tube stable. Sonically it should be punchier in the bass.

RC smoothing can be very quiet at high frequencies (hiss) but power line variations can show up in the output of the mic, as I wrote in an earlier post. That is noise, just not hiss, but subsonic garbage nonetheless. If you're lucky it will be very low level. In addition, the current draw of the amplifier, being AC, will interact with the passive power supply. A cap's impedance goes up as the frequency goes down thus the power supply becomes part of the amplifier stage.

Sonically I prefer to work with everything well regulated. In multistage amplification Isolating the B+ for each stage provides stable imaging keeping the instruments where you put them in the mix.

The buzz may be your power supply has poor grounding. Go to a negative regulated filament supply. A LM7906 is a good choice followed with a 1000mfd cap, bypassed with a 1 mfd mylar cap. If that's not quiet enough you can do double regulation.

Both B+ and Filament share the same ground wire in the cable, which means one can modulate the other if there are varying ground currents. The way around that is either send independent grounds to the mic or tie the grounds together at the output of each power supply. Varying current causes varying voltage due to the resistance of the wire.

HOWEVER!!:
That said, I need to see a schematic of your mic with parts values and voltages and see how the tube is used, please. I need to know how many milliamps of B+ the mic will draw.

 

[rock soderstrom]

The advantage of a regulated power supply is stability and power line immunity.

One advantage is that you can adjust the heating voltage to the existing transformer. In addition, you can easily adjust the sweet spot of the heater, keyword underheating experiments (whereby I would leave it at the original 6.3V)

Sonically I prefer to work with everything well regulated. In multistage amplification Isolating the B+ for each stage provides stable imaging keeping the instruments where you put them in the mix.

For microphones, many prefer unregulated B+ for sound reasons, or perhaps just because that's how the historical models did it. I have no clear opinion on this, because I have made too few direct comparisons and "sound" is subjective anyway.

The buzz may be your power supply has poor grounding. Go to a negative regulated filament supply. A LM7906 is a good choice followed with a 1000mfd cap, bypassed with a 1 mfd mylar cap. If that's not quiet enough you can do double regulation.

The buzz can also come from the previous modifications or numerous other reasons. Check everything again, Harmony!
The 7906 will not be good enough, because the negative grid bias is obtained from the heater, thus all noise components will be greatly amplified. The LM317 with additional filtering as described by Ruud is better.



This is how it was done with the original U67.

1Gen:


2Gen.

 

[rock soderstrom]

That said, I need to see a schematic of your mic with parts values and voltages and see how the tube is used, please. I need to know how many milliamps of B+ the mic will draw.

0,6-1mA

 

[Mr. Moscode]

Gracias. The schematics have all you need for the filament supply.
Since the fil winding on the xfmer is isolates just convert it to a neg supply. Use a 7915 neg VR and an RC filter to bring it to a quiet 6.3. Use the R4, R6, & R8 values as in the original circuit. Do not follow R8 with a large cap as it could charge to 15 volts if the PS is on and not connected to the mic. R8 limits the current to the filament to keep it from burning out in that scenario.

Will get back to you with a good hv reg.

 

[Khron]

... Or follow that with a capacitance multiplier...

 

[Mr. Moscode]

YUP.

 

[Mr. Moscode]

Line Isolated passive 210 Volt regulator.

You should get some very quiet DC with good line voltage isolation. Rejection will be around 100 db, no hum.

The MPSA92 pnp hv xstor, Led, zener string and resistors form a hi volt constant current source that feeds the zener string and isolates line voltage variation, keeping the zener current and operating point constant. The zener current is around 1ma which will make the zener (s) dissipate 220V x .001 = .22 watts. You can add zeners in series together to get your required voltage if needed and that will also distribute the power dissipation.

The mosfet is a 800v n-channel used as a source follower. The 1 meg R and 4.7 mfd cap connected to the zener filter zener noise from the gate. The 10k gate R is an oscillation stopper. The 1n4007s protect the mosfet from reverse current bleeding. The mosfet will dissipate 80V x .003A = .24 watt, well within the safe area of a TO220 mosfet rating of a watt, free air.

The bleeder is the 100k 2w R and keeps a load on the regulator to prevent a rise in unloaded output voltage. The caps and RC net finish off the noise rejection. Since plate current is .85 ma load regulation is not included although the bleeder helps as shunt regulation.

If this PS were to run multiple mics another output filter RC net could be added for isolation.

AS with any outboard power supply connect the load before powering it up and let it bleed down before unplugging the load.

This may be more than you need. I tend to do overkill. Bass notes should have good impact.