V76 Sowter choke question

[tomas.borgstrom]

I just finished a unit with two V72 (left) and one V76 (right). They sound great and are very quiet. I've compared them to original units and I think the sound quality is just as good. I actually like my V76 more since I've removed the filters and therefore get a more open sound. Before I get to the question I would like to brag about the complex switch I built for the V72s. It allows me to switch between the circuits of V77, V72, V72S, V72S EMI and V78 (I know they have different transformers etc). Besides the gain flexibility the sound changes when I switch. I like them all. The clearest difference is V77. Without the positive feedback the sound is a lot cleaner and the top sounds more open.

Now to my question. The voltages of the V76 match the schematics very close except for the plate where it's about 40-50V higher. The only conclusion I can make is that the Sowter chokes are not perfect clones of the original (or the values on the schematics are wrong). Would you leave it as it is or add a resistor in series before the choke? In other words. Whats the most important? Correct plate voltage and bias or avoid the increased distorsion and altered load of the tube with the added resistor?

 

[Dwarf Mc'Dougal]

Hi Tomas,
Sounds like the DCR of the Sowters isn't the same as originals. Not a big deal as long as inductance is high enough.
I think the V76 original chokes have DCR of about 18K and 7-8K respectively but it's been a while since I measured them.

Anyway, the bias of pentodes is set by the screen grid voltage, so if the screens are about where they should be then the bias will be fine.
Which choke loaded plate is higher, tube no. 2 or tube no. 4?
The E83F (no.4) has a max plate voltage of some 210V so if it's higher than just over 200-210V then I'd be inclined to add a resistor before the choke to bring it down.
If the E83F screen grid is at about 140V then the plate current will be about 10 - 11mA with the cathode resistor as shown, so just use ohm's law for your resistor value. Screen will be about 1/5th the current of the plate fwiw

Tube no. 2's anode, even if high, should still be within the limit of the tube and it would be your call on whether to fudge it down or not.
A higher plate V will increase the tube's gain, but since both stages (1+2, and 3+4) have loop feedback this translates into more feedback voltage being applied when the loop is closed. Stability may be an issue here, but maybe not. You'd want to check this to be sure.
Hope that helps.
Nice job

 

[tomas.borgstrom]

Hi Tomas,
Sounds like the DCR of the Sowters isn't the same as originals. Not a big deal as long as inductance is high enough.
I think the V76 original chokes have DCR of about 18K and 7-8K respectively but it's been a while since I measured them.

Anyway, the bias of pentodes is set by the screen grid voltage, so if the screens are about where they should be then the bias will be fine.
Which choke loaded plate is higher, tube no. 2 or tube no. 4?
The E83F (no.4) has a max plate voltage of some 210V so if it's higher than just over 200-210V then I'd be inclined to add a resistor before the choke to bring it down.
If the E83F screen grid is at about 140V then the plate current will be about 10 - 11mA with the cathode resistor as shown, so just use ohm's law for your resistor value. Screen will be about 1/5th the current of the plate fwiw

Tube no. 2's anode, even if high, should still be within the limit of the tube and it would be your call on whether to fudge it down or not.
A higher plate V will increase the tube's gain, but since both stages (1+2, and 3+4) have loop feedback this translates into more feedback voltage being applied when the loop is closed. Stability may be an issue here, but maybe not. You'd want to check this to be sure.
Hope that helps.
Nice job

Thank you for a very informative answer.

When I look at different designs the voltage drop / plate resistance is always higher when there's a resistor. Any explanation for this? Is there less need for voltage swing when you use a choke?

I've also built a V41 and I'm currently experimenting with the plate chokes/resistors. In the schematics there's 120V drop over the resistor to the plate of the first tube, no voltage drop over the choke to the second tube and only 10V voltage drop over the choke to the third tube. Both the second and third EF12 are used as triodes so in I guess in this case the plate voltage is more important for the correct bias.

 

[dmp]

When I look at different designs the voltage drop / plate resistance is always higher when there's a resistor. Any explanation for this? Is there less need for voltage swing when you use a choke?

That is exactly why a choke is used instead of a resistor. A choke is a more expensive part, so why use it? The DCR is lower, so the tube biases to a higher plate voltage. The choke has impedance to AC signals, so when amplifying, the choke makes the plate voltage swing and delivers higher output in comparison to a resistor. So with a choke, the tube stage can deliver higher headroom.

The only conclusion I can make is that the Sowter chokes are not perfect clones of the original (or the values on the schematics are wrong). Would you leave it as it is or add a resistor in series before the choke?

The tube you are using could also have variability. Measure the voltages upstream of the choke and measure the voltages of the tube pins. Ohm's law can tell you a lot about what is happening (the schematic you posted above has the info necessary to calculate the choke DCR, R = V/I).
You'd like to know how many mA the tube is conducting along with the DCR of the choke. I'd be surprised if the Sowter DCR is far off from the originals, as it would mean the wire AWG or turns is different. I used chokes wound by CJ for a V76 build and posted some info here

 

[Dwarf Mc'Dougal]

Thank you for a very informative answer.

When I look at different designs the voltage drop / plate resistance is always higher when there's a resistor. Any explanation for this? Is there less need for voltage swing when you use a choke?

In actual fact, the AC voltage is allowed to swing much more with a choke plate load versus a resistor plate load.
With a resistor, the swing can only be between the B+ value down to cathode.
But with a choke, energy is stored in the magnetic field and we can get voltage swing up to twice the voltage at the plate.
In the case of your E83F with about 200V on its plate, this would be equal to having a B+ value of about 400V and using a resistor that drops half of that in the typical manner.

 

[tomas.borgstrom]

Ohm's law can tell you a lot about what is happening (the schematic you posted above has the info necessary to calculate the choke DCR, R = V/I).

Thank you so much for making things clearer for me.

I use ohms law to calculate the plate current but the schematics says 200V on both sides of the choke and 2mA across it. This must mean that the DCR is very low.

I know that the choke with 7mA current is 65 Henry at 1 kHz and a DCR of 3.3 kOhm. In theory that would result in a voltage drop of 23V not 10V.

Is it desirable for a choke to have as high inductance and as low DCR as possible?

 

[dmp]

I use ohms law to calculate the plate current but the schematics says 200V on both sides of the choke and 2mA across it. This must mean that the DCR is very low.

For the V76, using the schematic you posted above, for BV515:
Vplate=170v. The B+ at upstream node is 279.3V (using Ohm's law with the mA values provided on the schematic)
With I=14.6mA, the DCR = 7.486 kOhms

For BV514:
Vplate = 175V. The B+ upstream node is 242.7 V
With I=3.6mA, the DCR = 18.8 kOhms

For the V41b, the schematic seems to be giving you 'rough' voltages. And the current amounts do not make sense. You cannot have 200v on both sides of the choke with 2mA current. At the top, the choke DCR is given as 3k3 ohms, which would be 6.6V drop across the choke. There must be that much voltage drop across the choke for the DC operating point. And the total labelled current to the plates of both tubes is much less than the current required to drop 30v across R37.

 

[MaxxDM]

Without the positive feedback the sound is a lot cleaner and the top sounds more open.

That's odd, because, the positive feedback in a true v72 makes the amp faster-sounding, so brighter as well

The V77 sounded a bit darker (I've had both for a few years), although part of that was also the input transformer, which had a lower impedance and higher ratio, If I remember correctly.

The Sowter TX are not exactly the same as the originals, so best to adapt the circuit, to optimize, I suppose.

 

[emrr]

Sowter reproductions always have significantly lower DCR (larger wire) than originals. Any repro. Inductance may be the same but voltage drops will always be less.

Vintage transformers almost always use significantly finer gauge wire than anyone uses today. Inductance is also usually much much higher in vintage transformers.

 

[tomas.borgstrom]

That's odd, because, the positive feedback in a true v72 makes the amp faster-sounding, so brighter as well

The V77 sounded a bit darker (I've had both for a few years), although part of that was also the input transformer, which had a lower impedance and higher ratio, If I remember correctly.

The Sowter TX are not exactly the same as the originals, so best to adapt the circuit, to optimize, I suppose.

My impression from the original units is that the V77 is more open with less colour. I had a pair some years ago.

 

[MaxxDM]

My impression from the original units is that the V77 is more open with less colour. I had a pair some years ago.

That’s odd.

I had 8 V72 that sounded pretty much identical once I replaced the psu caps, as well as the mp cap on the output, as well as all NOS tubes that I got in Italy.

The pair of V77 were given the same treatment, and they sounded less fast to me.

The input TX was for dynamic mics, and it loaded my tube condensers and 4038 mics a bit more.

In general, positive feedback increases slew rate. Too much and it creates instability, of course.

perhaps it’s a matter of how everyone hears things differently.

I think on something like drums, there positive fb should increase snappyness

I don’t have any of those units anymore, so I am going by my deductions of 15 years ago.

 

[Dwarf Mc'Dougal]

That’s odd.

I had 8 V72 that sounded pretty much identical once I replaced the psu caps, as well as the mp cap on the output, as well as all NOS tubes that I got in Italy.

The pair of V77 were given the same treatment, and they sounded less fast to me.

The input TX was for dynamic mics, and it loaded my tube condensers and 4038 mics a bit more.

In general, positive feedback increases slew rate. Too much and it creates instability, of course.

perhaps it’s a matter of how everyone hears things differently.

I think on something like drums, there positive fb should increase snappyness

I don’t have any of those units anymore, so I am going by my deductions of 15 years ago.

My findings agree with yours Maxx.
The input transformer used in the V77 presents a lower load to microphones and seems more suited to dynamic types.

I like the DC feedback (negative and positive) in the V72, it's a very elegant circuit.

Interesting aside: Attached is a picture of a large but portable (Ha!) tube mixer which I believe recently sold or else it's been taken off the market.
It's one of two made for Deutsche Gramaphone in 1959.

All mic inputs are standard 34dB V72's with the 4.X V76S amps used solely for the 4 bussing amplifiers.

Sorry the pics are screenshots btw.

 

[tomas.borgstrom]

My findings agree with yours Maxx.
The input transformer used in the V77 presents a lower load to microphones and seems more suited to dynamic types.

I like the DC feedback (negative and positive) in the V72, it's a very elegant circuit.

I'm not really sure what we're arguing. I actually like V72 more than V77. I sold my V77 pair some years ago because I almost never used them. I agree that V77 work best with dynamic microphones. My main use was on drums with MD421. In that context I found that they sounded more transparent and airy than other preamps. My impression and memory can be wrong but I recall V72 to sound richer and more "gritty" but not fast and snappy. I actually think that both are a bit overhyped and that V76 is much better, which I guess most will agree on.

So why did I then build a pair. Because I really enjoy building tube amps and I already had most of the parts (partly due to Drip not being able to select the correct components for their projects). I'm currently really happy with my build and the idea with the switch. The real test is always how much you still use the unit after a couple of years. From a theoretical point I can't say why the V77 circuit sounds cleaner in my build. I will try to do some recordings that highlights the differences.

While we're discussion the sound of the V series I can add that I just serviced four V76m. I was thrilled and brought them to a recording session. I was super disappointed. They sounded thin and uninspiring on every source and in the end I just went with my standard workhorses API, Neve, Crane Song etc.

 

[Dwarf Mc'Dougal]

I'm not really sure what we're arguing. I actually like V72 more than V77. I sold my V77 pair some years ago because I almost never used them. I agree that V77 work best with dynamic mics...

I'm sorry if I gave the impression I was in any way arguing with anyone. I was simply agreeing with MAXX about the lower load from the V77 input transformer which, since you used a modern type transformer in your own build, won't have the same issues anyway.

On to your question at the start: I do think that Sowters inductors will likely not have the same DCR as original design.
Again, not a massive deal as long as within tube limits and the units work.
If you can measure the screen grid voltages and cathode voltages for us, as well as raw B+ and the pre and post voltages of the two inductors, I'm sure we can put your mind to rest.

 

[tomas.borgstrom]

I'm sorry if I gave the impression I was in any way arguing with anyone. I was simply agreeing with MAXX about the lower load from the V77 input transformer which, since you used a modern type transformer in your own build, won't have the same issues anyway.

On to your question at the start: I do think that Sowters inductors will likely not have the same DCR as original design.
Again, not a massive deal as long as within tube limits and the units work.
If you can measure the screen grid voltages and cathode voltages for us, as well as raw B+ and the pre and post voltages of the two inductors, I'm sure we can put your mind to rest.

Haha, my mind is at rest. You have provided me with great answers. I can't measure it now because I'm currently adding a NY Dave preamp in the small space that is left inside the case.

I'm trying to get picture of what sound to expect from different designs and which units that are textbook examples of tube engineering.

A friend of mine who know a lot of electronics said that the engineers at Institut für Rundfunktechnik must have laughed during the coffee break when they saw the schematics of REDD47...

 

[MaxxDM]

I used to have a v76s but thought the v72 sounded less distorted. It had new tubes and all the electrolytics (lots) were replaced.

just shows you how everyone hears things differently.

I think it was more useful, in the sense that it was instantly adjustable

 

[CJ]

The original v76 chokes used a three chamber bobbin and Sowter uses a two chamber.

The problem with the high turns chokes is capacitance. If it is too high, it will shunt your high end into the filter caps. So your audio chokes are different from your power supply chokes, the latter only concerned with 60 or 120 Hz which is pretty low compared to 20 KC.

The number of chambers versus capacitance is quadratic.

C= 1/n^2

So C = 1 for one chamber

C= 0.25 for two chambers

C = 0.111...for three chambers

So the difference,

0.111/0.25 = 0.444...

So the originals have 44 percent less capacitance than the Sowter chokes assuming equal turns and similar core size.

Does this make a difference in sound? You would have to A/B it to find out. The final tube might have enough power to charge those little capacitors hiding out in the winds, without losing HF drive power to the output transformer.