Cheap 251 Using Russian 6S6B-V valve and other leftovers. Noisy results!

[tomas.borgstrom]

I've tried to lower the noise in tube microphones by increasing the grid resistor fx from 30M to 500M but I can't measure or hear any difference. The noise of the tube seems to be the most determining factor. When the resistor value gets above 500M I find that the microphone sounds worse. Like it's choked in some way.

 

[Tubetec]

Maybe the effect of the incidental capacitances around the hi-Z area are more pronounced with higher values of grid resistor , so you loose a little of the edge .
Seems like the best grid resistance value might found by adjusting on test and ear .

Its hard to quantify a feeling you get off a piece of audio equipment , but when you compare one to another and it jumps out and grabs you by the ears thats something that can be hard to pin down with test signals , a screen or digital readouts .

 

[Somkereki]

I was bored so I decided to go though my leftovers to see what could be made. Came up with, guess what, another 251 type build done in a few hours to see if these Russian 6S6B-V tubes I bought made any sound. Yes they are a bit noisy, but I think my bias is off and the fact I used carbon grid/cap bias resistors did not help either. The other problem was the cheap capsule I put on it has a small tent so It really needs some more work to be functional. Any Bias advise for the 6S6B-V tube would be appreciated! See both my schematic and the original 251. I omitted the 100pF cap at the output. Is this really needed?
View attachment 95450
Blank Body
View attachment 95452
Board Installed
View attachment 95453
Voltage divider and capsule bias resistor placement
View attachment 95455
Grid dip bias and plate current resistor placements. Other lead is the cathode bias resistor and output cap underneath.
View attachment 95456
Cathode bias resistor placement
View attachment 95457
Cathode bias and output caps placed.
View attachment 95458
Tube placement
View attachment 95459
Capsule and output transformer placements. 7 pin XLR placed with ground and power leads soldered in place.
View attachment 95461
Capsule bias capacitor placed as well.
View attachment 95462
Finished other side.
View attachment 95463
Familiar Apex body
View attachment 95465
My circuit kludge. I don't like the wires to cross unconnected like the old schematic!
View attachment 95466

Hello ! The noise is probably really from the carbon resistors used, because I've converted quite a few Neumann KM-56 and UM-57 mics to this tube, but the noise is reduced compared to the original AC701 and EC92 tubes. I assembled the necessary resistors of several 100 Mohm from low-noise 47 Mohm 0508 SMD resistors (impregnated in vacuum with epoxy resin produced for high voltage insulation purposes. I also replaced the other resistors with low-noise metal layer resistors. I also checked the capacitors with a Radiometer IM6 Megohmmeter. (Variable test voltage 1 It measures between -999 volts and up to 1000 Teraohms.Unfortunately, I have already encountered noise caused by the leakage current of the 1nF styro capacitor coming from the capsule.Good luck!
"Per aspera ad astra"

 

[Papa Tango Charly]

Great, thanks.
Way better noise for 6s6b in this example. I can see there is some compression going on with 6s6b tube by waveforms. Still sounds mellower. Fuller low end with 6s6b.

6072 still cleaner, more transparent. No winner here if you ask me, just different characters. Anyways wonderful piano sound with both. (I am a piano tuner)

For guys not hearing the difference, you need to split the files, lay them into own tracks, align them perfectly, and match levels as close as possible by phase cancelling. Select 0.5-1 sec loop and play in various parts by switching back and forth between files. If not having treated room use headphones, room rumble can mask difference in low end.

These are very subtle differences, not easy to spot them just by playing files one after another. We have very short hearing memory, and it can be very deceiving. I like to start listening to shootouts like this imagining i am half deaf, that i have no experience what so ever, and use technology to help me as much as possible. I have learned one thing, our hearing is very deceiving and fragile.

In my circuit with B+ 120v i had to use 50k plate resistor with 6s6b as i didn't like it running at about 40vdc with 100k. With 50k it runs at about 80v, didn't measure current but chinese psu seems to handle it well.

So you used a big core transformer in your build. Ok i thought it was a t14/1!
Definitely there is some biasing in how i hear things in blind listening... Anyway, i think i will use your circuit value as starting points for 251 circuit. I like the results you have.

Well i have m49(b and c) kind of circuit already built too using 7:1 transfo (a sowter mic input transformer) and didn t experience loss of bass too. But there is feedback involved in both circuit i build so output z of stage is lowered and it change the way the transformer is loaded...so in a bypassed grounded cathode amplifier the ratio needs to be higher.
I do think that for 6s6b with one between 8 /10 to 1 it should be ok, even better if it is one big core/high nickel content because there is low end with this tube. This is what i think make the difference in transformer needed for 6n16 and 6s6b ( and i prefer minimum 10/1 for 6n16 too).

Anyway i don t have any transformer with his ratio at home for now so it ll be speculation until i can afford some to test.

Back to m49, this discussion make take a detail i overlooked in my build: interelectrode capacitance.
There is difference between ac701 and 6s6b too, so c4 should be modified in value to compensate for this if you want to have close behavior. I will in my next build of m49b!

What do i expect from mod in bias, well... differences but i can t tell what they will be without test! From my experience with grounded cathode when you increase current (up to a limit) there is more details coming through, probably due to more second harmonic distortion or a different ratio of the different overtone created.
The circuit usually do seems to be quieter too. But i must say most of the test i made about that was in conventional preamp stage, not in mic impedance converter. Here the design are a bit spooky in using the tube at low current/voltage and with so high input z and results may well be differents!
Overall higher bias current give for my taste (so subjectively )better results.

But Kingkorg comments about his observation seems to agree with that idea..

Helllo I'm a beginner would like to build this microphone "In my circuit with B+ 120v i had to use 50k plate resistor with 6s6b as i didn't like it running at about 40vdc with 100k. With 50k it runs at about 80v, didn't measure current but chinese psu seems to handle it well" I would like to built this microphone. i'm looking for help. thank you.

 

[kingkorg]

I have seen C12 made with 56K at the plate. Has anybody tried with these values and come up with a diagram that made a great mic with a 6S6B-V. This 6S6B-V is almost like a Fet i'm shure that this microphone could be a killer. B+ 120V and heater 6V 7:1 transformer GTZ. Let me know if you have any idea. I need inspiration from you guys.i'm a begininner i would like to build an simple microphone. I f you have a plan. Let me know. Thank you. Best

You can use a plate resistor that gives you roughly 1/2 of PSU voltage at plate. So if you use 120v PSU aim for about 60v. It's more complicated than this, but without going into a rabbit hole you can use this as a guideline. 100k works well most of the time.

 

[Somkereki]

Maybe the effect of the incidental capacitances around the hi-Z area are more pronounced with higher values of grid resistor , so you loose a little of the edge .
Seems like the best grid resistance value might found by adjusting on test and ear .

Its hard to quantify a feeling you get off a piece of audio equipment , but when you compare one to another and it jumps out and grabs you by the ears thats something that can be hard to pin down with test signals , a screen or digital readouts .

I agree ! Hi-Z is not the work area we are used to! Here, very small differences modify the phase sequence, which our ears are very sensitive to. Don't forget that if a circuit has an amplitude breaking point at 20 Kilohertz, the phase error starts at 5 KHz! I also made a pre-divider head with 10 GOhm resistance for measuring condenser microphones (for a Metix MX54C instrument, which has a separate measurement limit of 500mV, where I can read the z result in volts.). Previous experience: I received a Neumann UM57 microphone, which was particularly noisy. I noticed that someone had previously replaced the original polystyrene insulator (which connects the head to the grid of the triode) with PVC material. Unfortunately, with these old microphones, the original insulation sometimes cracks. He also replaced the cable from the connector to the diaphragm with a PVC cover. I turned a polystyrene insulator into the connector and replaced the wire with PTFE insulation (silver coated). The popping noise is gone! After that, of course, I replaced the 500 MOhm resistors and the rest with low-noise metal layer resistors. As a result, a signal / noise improvement of more than 20 dB can be achieved! We are all "trying to catch the tail of a black cat in a dark room" I directly made a switch for the A-B tests with a 4 Morse gold-contact SDS relay, which can be controlled remotely from the listening position. My monitor speaker is an Acoustic Research Spirit 142 pair. (modified) Sometimes I wonder how versatile the British company Teledyne is, since they also manufactured these, my Neutrik A2-D Audio Test & Service System instrument and my LeCroy WaweAce 204 digital storage oscilloscope. The rest are usually Tektronix and HP. I wish everyone good luck and a happy, peaceful Christmas !