Anyone built a sela t24/25 clone?

[RuudNL]

R6 = 100K ????

 

[Purplenoise]

R6 = 100K ????

Yes…actually I have 56k in my build. Isn’t that a little similar to the u47?

 

[RuudNL]

I would leave the anode resistor (R2) 100 K and lower the value of R6 to maybe 10K or so.
(A lower anode resistor value will reduce the gain.)
And make R4 = R5 for a polarisation voltage of 60 V (both 470 K or 1 M.)
But of course there are different ways to make things work!

 

[Purplenoise]

I would leave the anode resistor (R2) 100 K and lower the value of R6 to maybe 10K or so.
(A lower anode resistor value will reduce the gain.)
And make R4 = R5 for a polarisation voltage of 60 V (both 470 K or 1 M.)
But of course there are different ways to make things work!

Great, thank you so much for the suggestion. Really appreciate it. Just for educational purpose…why is it better like that? Isn’t the anode resistance additive? 56k + 56k as opposed to 100k+10k?
( Yes I have both polarization resistors at 1M)

 

[RuudNL]

An anode resistor of 100 K. gives a reduced load for AC voltages. (Read: the output signal of the tube.)
The second 100 K.resistor with the 2 µF capacitor only forms a low-pass filter. With a good power supply this shouldn't be necessary. But what is worse, is that you lose as much (DC) voltage over the 'filter' resistor as in the anode resistor. This reduces the voltage swing on the anode considerably.
Usually, you want the anode voltage at about 1/2 of the supply voltage, so that the positive swing can be as large as the negative swing.
Suppose you have an anode current of 0.5 mA, then the anode voltage will be:
120 - (0.5 mA x 200 K) = 20 V. There is now 50 V. over the filter resistor and 50 V. over the anode resistor.
The 'supply voltage' at the filter capacitor is 70 V.
(An anode current of 0.5 mA is already a very low value, if you would use a higher anode current, the situation would be even worse!)

If we reduce the value of the filter resistor to for example 10 K, we get the following situation:
Anode voltage is now 120 - (0.5 mA x 110 K) = 65V. There is now 5 V. over the 'filter resistor' and 50 V over the anode resistor.
The 'supply voltage' at the filter capacitor is now 115V.
The voltage swing on the anode can now (theoretically) be between 0 V (fully conducting) and +115 V ('open')
This is a lot more than the 70 V we had before (supposed the tube is correctly biased).
Of course we don't need an output of 70 V peak-peak, but in reality there is a step-down output transformer of maybe 10:1 and that reduces the output already to 7 V pk-pk. (Or 2.47 V rms)

The moral of this story: don't waste valuable DC voltage over the 'filter' resistor, you want the highest voltage to feed the anode circuit for maximum undistorted SPL.

 

[Purplenoise]

An anode resistor of 100 K. gives a reduced load for AC voltages. (Read: the output signal of the tube.)
The second 100 K.resistor with the 2 µF capacitor only forms a low-pass filter. With a good power supply this shouldn't be necessary. But what is worse, is that you lose as much (DC) voltage over the 'filter' resistor as in the anode resistor. This reduces the voltage swing on the anode considerably.
Usually, you want the anode voltage at about 1/2 of the supply voltage, so that the positive swing can be as large as the negative swing.
Suppose you have an anode current of 0.5 mA, then the anode voltage will be:
120 - (0.5 mA x 200 K) = 20 V. There is now 50 V. over the filter resistor and 50 V. over the anode resistor.
The 'supply voltage' at the filter capacitor is 70 V.
(An anode current of 0.5 mA is already a very low value, if you would use a higher anode current, the situation would be even worse!)

If we reduce the value of the filter resistor to for example 10 K, we get the following situation:
Anode voltage is now 120 - (0.5 mA x 110 K) = 65V. There is now 5 V. over the 'filter resistor' and 50 V over the anode resistor.
The 'supply voltage' at the filter capacitor is now 115V.
The voltage swing on the anode can now (theoretically) be between 0 V (fully conducting) and +115 V ('open')
This is a lot more than the 70 V we had before (supposed the tube is correctly biased).
Of course we don't need an output of 70 V peak-peak, but in reality there is a step-down output transformer of maybe 10:1 and that reduces the output already to 7 V pk-pk. (Or 2.47 V rms)

The moral of this story: don't waste valuable DC voltage over the 'filter' resistor, you want the highest voltage to feed the anode circuit for maximum undistorted SPL.

Got it, thank you so much for taking the time to explain this. I will definitely give it a try. Really appreciate it.

 

[Purplenoise]

An anode resistor of 100 K. gives a reduced load for AC voltages. (Read: the output signal of the tube.)
The second 100 K.resistor with the 2 µF capacitor only forms a low-pass filter. With a good power supply this shouldn't be necessary. But what is worse, is that you lose as much (DC) voltage over the 'filter' resistor as in the anode resistor. This reduces the voltage swing on the anode considerably.
Usually, you want the anode voltage at about 1/2 of the supply voltage, so that the positive swing can be as large as the negative swing.
Suppose you have an anode current of 0.5 mA, then the anode voltage will be:
120 - (0.5 mA x 200 K) = 20 V. There is now 50 V. over the filter resistor and 50 V. over the anode resistor.
The 'supply voltage' at the filter capacitor is 70 V.
(An anode current of 0.5 mA is already a very low value, if you would use a higher anode current, the situation would be even worse!)

If we reduce the value of the filter resistor to for example 10 K, we get the following situation:
Anode voltage is now 120 - (0.5 mA x 110 K) = 65V. There is now 5 V. over the 'filter resistor' and 50 V over the anode resistor.
The 'supply voltage' at the filter capacitor is now 115V.
The voltage swing on the anode can now (theoretically) be between 0 V (fully conducting) and +115 V ('open')
This is a lot more than the 70 V we had before (supposed the tube is correctly biased).
Of course we don't need an output of 70 V peak-peak, but in reality there is a step-down output transformer of maybe 10:1 and that reduces the output already to 7 V pk-pk. (Or 2.47 V rms)

The moral of this story: don't waste valuable DC voltage over the 'filter' resistor, you want the highest voltage to feed the anode circuit for maximum undistorted SPL.

Tried it…man it actually works really well. Opened up the sound nicely and made it more even. This is the best combination so far with this tube (5718a). Surprisingly similar sounding to my ptp 47 “clones”.
I think this mic is done haha…time to make a second one.
Thank you so much again and everyone else for all the help and priceless info.

 

[doqmemory]

Tried it…man it actually works really well. Opened up the sound nicely and made it more even. This is the best combination so far with this tube (5718a). Surprisingly similar sounding to my ptp 47 “clones”.
I think this mic is done haha…time to make a second one.
Thank you so much again and everyone else for all the help and priceless info.

Now I want to make one, so I guess this will get one of my 5718's.

 

[Purplenoise]

Now I want to make one, so I guess this will get one of my 5718's.

I really like it so far. I have a dachman da-k47 ( probably 3u audio made) and sounds really really good. The transformer is a cinemag 2510 8:1. Maybe a bv11 10:1 will work even better.

 

[tomas.borgstrom]

300k resistor replaced by 1M to get normal 60v polarization instead of 92v and plate resistors changed to 56+56k.
Transformer is a cinemag cm-2510 8:1ratio.

Why did you change the plate resistors?

 

[Purplenoise]

Why did you change the plate resistors?

Originally I planned on using a 6s6b so I left those plate resistor values when I installed a 5718. Since then, and after RuudNL’s suggestion, I switched back to 100k for plate and 12k for the filter. It sounds much better now

 

[Purplenoise]

So as I am playing around with this mic I changed the plate resistor to 180k leaving the filter resistor at 10k…I guess similar plate resistance to original schematic but instead of 100+100k now it’s 180+10k.
Not drastically different sound but a little cleaner low end and now the waveform on loud transients is perfectly symmetrical so I guess less clipping/distortion?

 

[Purplenoise]

Ok finally after a lot of tweaking I believe I have arrived to something I can find a lot of uses for. Needless to say that it’s not a t24 clone anymore but it’s own unique thing.

I settled on a k67 style capsule after trying a couple k47s and an m7. Some might like this sound ( it was a cool sound) but always sounded a little “muffled “ to me. Maybe that’s what happened in the original design and raised the pol voltage to 92v.
K67 works nicely though and in my case I reduced the 10k peak with an 680pf plate to ground cap. So far here are the parts/voltages/values.

-K67 capsule ( Chinese..kinda scooped sounding but nice)
-5718 tube
-110v B+
-5.9v H+
-cinemag 2510 8:1 transformer
-r2 180k, r6 12k, r4/5 1M
- 680pf plate to ground styrene de-emphasis cap
-electrolytic c3
-epcos mkt c1
-wima 0.01uf tff c2

The type of c2 and c1 has an influence on the sound ( it does) so it is worth experimenting.

 

[tomas.borgstrom]

So as I am playing around with this mic I changed the plate resistor to 180k leaving the filter resistor at 10k…I guess similar plate resistance to original schematic but instead of 100+100k now it’s 180+10k.
Not drastically different sound but a little cleaner low end and now the waveform on loud transients is perfectly symmetrical so I guess less clipping/distortion?

Glad to hear that you're happy with the microphone (that in the end became a new design).

If we stick with building a clone, are you sure that you can calculate plate resistance like that when you have a cap to ground in between?

Since you're using another tube a different plate resistance might work better. Keeping the filter intact makes more sense.

I'm currently not at home but I had a quick look in the Sela last night. The schematic in the start of this thread is correct. The grid voltage must arise from the capsule leakage. Next week I will post some pictures, audio and basic data.

EDIT: The grid voltage does not arise from capsule leakage. The circuit is grid leak biased. Please see post from mhelin on the next page.

 

[Purplenoise]

Glad to hear that you're happy with the microphone (that in the end became a new design).

If we stick with building a clone, are you sure that you can calculate plate resistance like that when you have a cap to ground in between?

Since you're using another tube a different plate resistance might work better. Keeping the filter intact makes more sense.

I'm currently not at home but I had a quick look in the Sela last night. The schematic in the start of this thread is correct. The grid voltage must arise from the capsule leakage. Next week I will post some pictures, audio and basic data.

Yeah I keep fine tuning “by ear” but this design is an easy start with great potential I think. So far it sounds kinda “modern” but smooth with very controlled sibilance…I have a feeling that a pair would be great on overheads but also guit cabinets. Not a very “natural” sound but kinda processed and mix ready.
Looking forward to your findings on the originals.

 

[tomas.borgstrom]

I finally got some time to explore the Sela T25. The build quality is good. It's very light and as small as a microphone with a K47, a tube and a transfomer can be, 114 x 41 mm. It's a two layer mesh, outer about 2.5mm, inner very dense. See the attached schematics for voltages. I don't have a voltmeter that can measure at impedances as high as 1G (I've planned to buy or build for several years). The transformer is 1:6.5 custom made by Sela. The photos are with stock components. Before the recordings I replaced the electrolytic capacitors with film. Since the T25 have the same capsule and tube as M49 (with K47) I recorded both with the guitar on a stand and the capsules at the exact same spot. It doesn't really work to record both at the same time since the distance between the capsules results in a substantial difference in sound. You have to tilt the Sela microphone quite much for the capsule to point forward.

Listen for yourself. Both sound really good but to me one sounds better. T25 has a really high output, about 10dB more than M49, and it is really quiet. When I lowered the capsule voltage to 60V the output dropped about 4dB. I expected larger differences between the two voltages. The M49 sound is often referred to as "velvety". Whatever that means I don't like that description, I think it has too much spark. Perhaps it's more correct if it is an M49b with M7. The recorded microphone is an M49c.

 

[tomas.borgstrom]

The uploaded clips doesn't seem to work. Here's a link to download them:

M49 and T25

 

[Purplenoise]

I finally got some time to explore the Sela T25. The build quality is good. It's very light and as small as a microphone with a K47, a tube and a transfomer can be, 114 x 41 mm. It's a two layer mesh, outer about 2.5mm, inner very dense. See the attached schematics for voltages. I don't have a voltmeter that can measure at impedances as high as 1G (I've planned to buy or build for several years). The transformer is 1:6.5 custom made by Sela. The photos are with stock components. Before the recordings I replaced the electrolytic capacitors with film. Since the T25 have the same capsule and tube as M49 (with K47) I recorded both with the guitar on a stand and the capsules at the exact same spot. It doesn't really work to record both at the same time since the distance between the capsules results in a substantial difference in sound. You have to tilt the Sela microphone quite much for the capsule to point forward.

Listen for yourself. Both sound really good but to me one sounds better. T25 has a really high output, about 10dB more than M49, and it is really quiet. When I lowered the capsule voltage to 60V the output dropped about 4dB. I expected larger differences between the two voltages. The M49 sound is often referred to as "velvety". Whatever that means I don't like that description, I think it has too much spark. Perhaps it's more correct if it is an M49b with M7. The recorded microphone is an M49c.

Tomas thank you so much for doing this. Really appreciate it you took the time. Looking forward to checking the clips after I am back from vacation.
How would you describe the sound of the t25…your personal take on it?

 

[LevinGuitar]

Very interesting info and audios, thank you!

Don't you want to try a C4 from M49c in the Sela?

 

[tomas.borgstrom]

Very interesting info and audios, thank you!

Don't you want to try a C4 from M49c in the Sela?

I believe the only thing C4 will do is to lower the output. Neumann selected that capacitor so the output would be the same in all microphones. It's not included in the early schematics. Does anyone experience that it affects the sound?