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

[SparkleBear]

Hey all, I tried to do some maths to calculate EF86 B+ load resistor to test/design a modded power supply with out endangering my tube. What is the proper way to figure this out? I just did some ohms law with R = 210v / 0.1A . Giving me max 21k. Is this correct?



I'm going to mod this Alctron psu for my u67 and want to know what the load would be on B+.



I just RE'd the circuit they are using and will fill in the values I need with the correct resistors.... maybe add more filtering if necessary. Potentially also make the H+ side non regulated?? Basically was trying to mod this like Chunger does for MK67 build thread but my pcb ended up being a different model... so I have to figure it out on my own. (which at first was a drag but now am enjoying the learning process.) I did get enough of the mod to remove the diodes in D5 and D6 position and add a bleeder to ground.

Here is the schematic i came up with from tracing the PCB.



I think I can take out the LM317 from the situation and use the resistor rails as a voltage divider to get from the 20vdc down to the 6.3vdc that I need for the tube (6.15v ideally).
Now I just need to know the B+ load for ef86 and I can hopefully calculate the correct resistors. This is my first time building/modding a PSU from scratch... so this is exciting.

I already built my DNU67 so modifying this is more of a research project to see if the "better" power supply really makes a difference for audio quality.

Any advice for improvements?

Thanks!!!

 

[rock soderstrom]

for the heater R=U/I 6,3V/0,2A=31,5R

I just did some ohms law with R = 210v / 0.1A . Giving me max 21k. Is this correct?

No, your EF86 will never draw 0,1A on the B+ rail. This is big power tube territory. I guess 0,001A or 1mA is more realistic.

 

[SparkleBear]

.01A is 1mA... sorry! NOT 100mA woopsie.

 

[SparkleBear]

No, your EF86 will never draw 0,1A on the B+ rail. This is big power tube territory. I guess 0,001A or 1mA is more realistic

So it is be 210k ohms = 210v / .001a.

With that value I should be able to do the math for the rest of the resistors.

 

[rock soderstrom]

.01A is 1mA... sorry! NOT 100mA woopsie.

Sorry, .01A is 10mA

 

[rock soderstrom]

So it is be 210k ohms = 210v / .001a.

With that value I should be able to do the math for the rest of the resistors.

Yes, try 210k and see what happens.

 

[SparkleBear]

Yes, try 210k and see what happens.

Sweet! So R1 - R4 are 12k and R6 is 16.5k (2x33k in parallel). Then the 25k trim pot gets me in the zone with a 210k B+ load. YES!

So now to work on un-regulating the H+ supply.

Do you see any other improvements i can make to the circuit? More filter caps? or different values?

Thanks !

 

[rock soderstrom]

Do you see any other improvements i can make to the circuit? More filter caps? or different values?

Later, when you have the mic connected, monitor the voltage when you turn it on, the polarization voltage can get very high when the tube is cold and no current is flowing.

You may need to clamp down the voltage for this moment with Zener diodes??

So now to work on un-regulating the H+ supply.

Member @RuudNL was succesful to mod a similiar PSU for U67 with regulation.

https://groupdiy.com/threads/heater-voltage-filtering.80421/post-1033597

 

[SparkleBear]

Well I have successfully modded this beast. Built a little filter board a-la @RuudNL. Thanks for your advice mate!

All thats left is to do the voltage clamp on the B+ output. I definitely see the voltage getting up to near 250v when turning it on. I took the capsule assembly off of the mic for initial testing... Also put in a beater tube. Don't want to damage my telefunken ef806s thats for sure.

Well... So for the zener diode... I was thinking of getting a 220v zener and strapping it between B+ and ground. Is that the right move?



So if the voltage goes above 220v the zener kicks on and creates a super low impedance path to ground so the voltage wont travel into the mic. Is that correct?

Thanks muchly!

 

[rock soderstrom]

Well... So for the zener diode... I was thinking of getting a 220v zener and strapping it between B+ and ground. Is that the right move?

There are two options from my point of view. As you described it or you use the Zener diode as a "regulator" and set it exactly to the B+ you want, here 210V.

There is some info about this here on Groupdiy.com

No matter what you do, wire a small capacitor (e.g. 0.1uF) in parallel to the Zener at the end of the PSU.

Edit: somewhere in this thread are some Infos about this topic AFAIK.
Thread 'Warm Audio WA-67 - Teardown' Warm Audio WA-67 - Teardown

 

[SparkleBear]

I wanted to go with an unregulated supply, at least for the B+ side. I like the way i described because the zener is out of the equation under normal operating conditions.

So far, this PSU does not sound as good as the DNU67 poctop psu. Its got a higher noise floor and the upper mids get congested compared to the DNU67... at least to my ears.

I am hoping to figure out how to filter or do something to make it sound better. It would be nice to have that little supply for traveling around and not having to lug a giant steel lunchbox ;P

 

[rock soderstrom]

So far, this PSU does not sound as good as the DNU67 poctop psu. Its got a higher noise floor and the upper mids get congested compared to the DNU67... at least to my ears.

Sound is relative, but the noise floor is measurable. Actually, your PSU is already well filtered. Is the noise floor slightly worse or significant?

 

[RuudNL]

Be careful with zeners, as they can create a lot of noise on the +HT line!
0.1µF won't help much, because if the zener is conducting, it represents a very low impedance.
Personally I would put the zener in the before last filter stage, so that the last filter stage can also reduce (filter) the zener noise. Of course you have to take the voltage drop in the last filter stage into account, so you probably will need a zener with a slightly higher voltage than you would have used if you had put it at the end of the power supply.
In a real circuit 210 or 220 V doesn't make a big difference!

Take care that if the power supply is for a real U67, you will need -6V for the filament (not +6v!)
This is because the (negative!) grid bias is derived from the filament voltage.
Of course you can use a LM317, but in this case you can't have a common ground for the filament voltage and the +HT. (As shown in the hand drawn schematic.) You are using different transformer windings, so it is easy to keep both (filament and +HT) outputs 'floating'.

 

[rmburrow]

.01A is 1mA... sorry! NOT 100mA woopsie.

0.001 A is 1 mA.

 

[rmburrow]

So it is be 210k ohms = 210v / .001a.

With that value I should be able to do the math for the rest of the resistors.

Typical anode resistor for the triode connected EF86 in a U67 with 210v HT supply is 150 k 1/4 watt film.

 

[SparkleBear]

Be careful with zeners, as they can create a lot of noise on the +HT line!
0.1µF won't help much, because if the zener is conducting, it represents a very low impedance.
Personally I would put the zener in the before last filter stage, so that the last filter stage can also reduce (filter) the zener noise. Of course you have to take the voltage drop in the last filter stage into account, so you probably will need a zener with a slightly higher voltage than you would have used if you had put it at the end of the power supply.
In a real circuit 210 or 220 V doesn't make a big difference!

So would a 250v zener be a better choice for the position? Or is the 220v okay? I definitely don't want the noise there. So like this?

Take care that if the power supply is for a real U67, you will need -6V for the filament (not +6v!)
This is because the (negative!) grid bias is derived from the filament voltage.
Of course you can use a LM317, but in this case you can't have a common ground for the filament voltage and the +HT. (As shown in the hand drawn schematic.) You are using different transformer windings, so it is easy to keep both (filament and +HT) outputs 'floating'.

Interesting!!! I do NOT have the filament ground separated from the main ground. I have not tried to power this up with a "real u67" yet. I have been using my MK67 tlm67 drop in with the russian ef86 tube for testing. I wouldn't want to risk my ef806s. So to float the heater i would isolate it from the rest of the main ground and just treat one of the AC windings as the "ground" ? How would you make sure the heater voltage is negative??

 

[RuudNL]

You could do it like this. (And still use a 'positive' voltage regulator)
See it like a battery: it doesn't matter if you ground the positive or negative side; it is 'a voltage and current source', completely separated from the rest of the circuit. But you may have to cut one or more traces on the PCB to separate the grounds!

 

[SparkleBear]

You could do it like this. (And still use a 'positive' voltage regulator)

Ahh i just am seeing this! I went on my own journey trying to figure that out!


I tried it with the filter board of your advice... but the voltage was jumping about wildly +/- 0.5v. Figured i needed to add more capacitors to the design. I just split down the resistor values and added another two capacitors. I need to get those parts before I can build. How does this look? Any improvements I can make?

The tube seemed to do OK with the positive heater voltage. What is the downside of it other than its not in the design?

 

[rock soderstrom]

Your schematic makes no sense to me.

The additional filtering for the LM317 should be on the output side of the regulator.

Similiar to this sketch from one of my projects.

 

[RuudNL]

I agree, this won't work!
Besides that, 4 x 10.000 µF is a kind of overkill after the LM317 regulator.
10 ohms in series and 1000 µF effectively kills all the noise.
(And I am sure about this, because I use this in my own U67 power supply.)