V76 tube preamp build

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

dmp

Well-known member
Joined
Oct 28, 2009
Messages
3,829
Location
Madison, WI
I started building this V76 a few years ago. Almost done.
Using a JT-115k-E90 transformer on the input and JT-10k61-1M on the output.
CJ wound the chokes.

 

Attachments

  • IMG_20180312_224440.jpg
    IMG_20180312_224440.jpg
    269.1 KB · Views: 377
The gain knob is 24 step, 2 dB/step
Has a pad switch for 0 dB, -24 dB, -36 dB, or -48 dB.
(originals had a three different pads incorporated with the gain switch -13dB, -24 dB, and -30 dB.

Gains:
with no pad:          30 dB to 76 dB gain
with -24 dB pad:    6 dB to 52 dB gain
with -36 dB pad:  -6 dB to 40 dB gain
with -48 dB pad:  -18 dB to 28 dB gain

A phase switch and DI on the left.


 

Attachments

  • v76_front.jpg
    v76_front.jpg
    90.4 KB · Views: 149
Thanks Ian

Some info on the chokes:

From the V76 schematic they should be:
V2: L=400H, DCR=18 kohm
V4: L=250H, DCR=9.3kohm

measurements using Instek LCR 821 measuring L and R
V2:
1V@100Hz:  415.9H, 31.13 kohm
1V@200Hz: 381.3H, 35.7 kohm
1V@400Hz: 378.8H, 55.7 kohm
1V@600Hz: 400H, 103.2 kohm
1V@700Hz: 442H, 195 kohm
1V@1000Hz: 520H, 406 kohm

V4:
1V@100Hz: 257H, 12.66 kohm
1V@200Hz: 245H, 13.7 kohm
1V@400Hz: 249H, 26 kohm
1V@600Hz: 262H, 52 kohm
1V@800Hz: 281H, 96 kohm
1V@1000Hz: 315H, 181 kohm

Here is a pick before I made the shielding cans.
 

Attachments

  • IMG_20180219_212433.jpg
    IMG_20180219_212433.jpg
    288.6 KB · Views: 219
I made steel cans for the chokes and added a layer of mumetal for the V2 choke. I also made a steel cylinder to surround the toroid power transformer and wrapped this with mu metal. I used one sheet of mu metal for screening  (EMI Shielding Metal - Ultraperm 80 'MuMetal', 10.5" x 8.0" x 0.004" from Noo elec, $22 per sheet). This is nice stuff - adhesive backed and soft enough that it can be easily cut with a scissors.
The shield around the input section is a piece of aluminum covered with adhesive backed copper foil. The shielding here is for RF interference, since there isn't a coil underneath to pick up magnetic field coupling.
The jensen input transformer I used came with -90 dB mumetal shield. The Jensen output transformer comes completely enclosed with a Faraday shield, per the datasheet.

About shielding, I reproduced some of DaveP's shielding tests. I found the same trend of decreasing magnetic coupling with distance and reduction of magnetic coupling with steel vs. aluminum in between. Placing a flat sheet between transformers is not going to be very effective at shielding, however - since the magnetic field doesn't just travel across in a straight line. Surrounding the transformers completely with a continuous shield is much more effective.

The original V76 was packed into a much smaller package, with the transformers right next to each other. The challenge to shield magnetic interference is much higher as the distance is smaller. Still, having the power transformer in the same box makes the build quite a bit more difficult.

The wikipedia article linked to below states: " The best shape for magnetic shields is thus a closed container surrounding the shielded volume. The effectiveness of this type of shielding depends on the material's permeability, which generally drops off at both very low magnetic field strengths and at high field strengths where the material becomes saturated. So to achieve low residual fields, magnetic shields often consist of several enclosures one inside the other, each of which successively reduces the field inside it."

Summary:
- Magnetic coupling decreases with distance according to an inverse square law (the magnetic interferance decreases inversely as the surface area of the sphere). This means if the distance between transformers is doubled (i.e. 50mm to 100mm), the interference decreases by 1/2^2, or 1/4. If the distance, quaddruples (50mm to 200mm) the interfence drops 1/4^2, or 1/8.
- The permeability of mumetal is much higher than steel and steel is much higher than aluminum.
- As DaveP showed, and the wikipedia article describes, several layers of magnetic shielding function better than a single thick layer.

The goal for magnetic shielding is to provide a continuous enclosure (i.e. Faraday Cage) that redirects the magnetic flux through the shield. This is explained well in section 1.2.5 in "audio transformers" by Bill Whitlock.
- The surrounding shield enclosing the transformer should be continuous (no air gaps).
- mechanical stress decreases permeability, so take care in shaping the shield enclosure to abuse it as little as possible, or re-anneal it.
- transformers with high permeability core (high nickel) will be more immune to pickup than lower mu, (steel cores)

The Jensen transformers use a high nickel (80%) core.

Something to note is shielding for RF and Magnetic coupling are both important in a preamp. Shielding magnetic coupling is important to isolate between transformers and chokes, particularly power and audio transformers. The power transformer will radiate 60 Hz to the input/output/chokes in the audio circuit. The output transformer can radiate to the input transformer and cause oscillation.
RF shielding on the other hand should prevent pickup of radio frequencies, i.e. prevent the input grid of the preamp circuit from acting like a radio. Shielded wire and copper shielding are effective at RF shielding.

RF vs Magnetic Shielding:
https://en.wikipedia.org/wiki/Electromagnetic_shielding

Audio Transformers by Bill Whitlock:
http://jensen-transformers.com/wp-content/uploads/2014/09/Audio-Transformers-Chapter.pdf

Magnetic Shielding:
http://www.magnetic-shield.com/faqs-all-about-shielding.html
 
Great work DMP,

I'm glad my tests were helpful.

Please let us know what noise levels you achieve

Best
DaveP
 
DaveP said:
Please let us know what noise levels you achieve
I'll try to post some measurements. I used to do measurements with RMAA but lost that setup so I will have to set something up again.
 
The gain switch in the original schematic had 12 positions of gain combined with the pad. There were only 7 positions of amplifier gain.
I calculated the resistor ladder in Excel. It is log scaled with a total resistance of 30k.
I lowered the last resistor to 56R from 98R to coax a little more gain out at the top end for kicks. Is stable with listening tests.
Used a nice elma switch on a PCB.
Here are the resistor values I used in kohms ( I took the closest value I had to what I calculated, and was able to keep it pretty close to 2 dB per step):

6.8000
5.6000
4.2200
3.3000
2.2000
1.8000
1.5000
1.1000
0.8200
0.6200
0.4700
0.3900
0.2800
0.2000
0.1600
0.1300
0.0910
0.0750
0.0560
0.0430
0.0330
0.0270
0.0200
0.056
 

Attachments

  • IMG_20180312_224318.jpg
    IMG_20180312_224318.jpg
    145.6 KB · Views: 123
I used Antek AS-05T320 - 50VA 320V Transformer. I used the 300 v winding and the 320v was taped off. One heater winding went to the indicator and the other went to the tubes.

Since this transformer is designed for 115v, my voltages were higher (121.7 V wall voltage). I brought them down with resistors. I used 0.22 2 watt resistors before the 100 ohm resistors connecting H+ to ground. This brought them to 6.2 v, when without they were ~6.6v. The EF804s tubes specify max +/-5% on heater voltages and I'd rather be a little low than high.

The B+ voltages were 284v and 241v  (schematic indicates 279v and 242v). I increased the 1k resistor to 6k (10k || 15k) to get them down to spec.

The tube plates voltages were:
        measured:              schematic:
v4   193                       170
v3   90                         80
v2   166                       175
v1   63                         67
 
So I did a quick measurement of the noise with the pad all the way up (-48dB) and the gain at max (78 dB). I measured 6 mV on my oscilloscope. I'll try to do some better measurements when I can get RMAA installed on our studio computer.

Listened to the noise in the studio last night and it was very clean - no noticeable hum or buzz. Just normal hiss.
 
That looks good to me. 6mV is about -42dBu which means with 78dB of gain your EIN is around -120dBu. Since you were looking on a scope you probably measured the peak noise so the rms will be at least 6dB lower than this which means you EIN is probably in the -126dBu region which is very good for a tube mic pre.

Cheers

Ian
 
My scope can measure RMS voltage and that is what I did, 6 mV Vrms.
I was thinking that was the way to calculate the noise figure, but it seemed unbelievably low.  But I did just see this on the historic info page at tab-funkenwerk:

"Two months after its introduction, the IRT ordered all radio stations to retube their V72’s and V76’s with the EF804S in order to bring down the self noise of the V72 to -88dB and -120dB for the V76. "

 
Front.
Not sure this is the best way to label the gain/pad, but I figure it will be the easiest to understand.
I thought about labeling 4 gain ranges with the pad positions.
 

Attachments

  • IMG_20180315_223803.jpg
    IMG_20180315_223803.jpg
    951.6 KB · Views: 139
Very nice.  I may have missed it, want to add it's worth going through 6-10 front end tubes after a 40 hour burn in to find the quietest. 
 
emrr said:
Very nice.  I may have missed it, want to add it's worth going through 6-10 front end tubes after a 40 hour burn in to find the quietest. 
Well I only have the 3 EF804s right now, but I can flip them around to see which gives the quietest output.
Unfortunately it is an expensive tube (~$50 each) but assuming they are NOS, they should last 10,000 to 100,000 hrs.
The ones I bought came in factory sealed boxes, so they are starting out fresh.
I had a little sputtering at first but after about 5 mins it went away. Would you recommend I leave it on for 40 hrs straight?
 
dmp said:
My scope can measure RMS voltage and that is what I did, 6 mV Vrms.
I was thinking that was the way to calculate the noise figure, but it seemed unbelievably low.  But I did just see this on the historic info page at tab-funkenwerk:

"Two months after its introduction, the IRT ordered all radio stations to retube their V72’s and V76’s with the EF804S in order to bring down the self noise of the V72 to -88dB and -120dB for the V76. "

Sounds like yours hits the spec dead on. Tube mic pres can do better than that. The limitation with the V72 and V76 is the first stage which is a pentode. A triode first stage would be quieter but even so you have a respectable performance.

Cheers

Ian
 
I should mention I left out all of the filter components in the section between v2 and v3.
I thought about putting them in, but I hardly ever use filters built into a preamp, so ultimately left them out.

With these out, you save the 6 dB of gain that's lost due to the filter section and the 80k resistor (which forms a divider with the 80k on v3 grid). So I put in a single blocking capacitor for everything between the plate of v2 and the grid of v3. The grid of v3 is still tied to ground with a 80k resistor.
I used a 1.5 uf, but anything above 0.5 would have been sufficient for putting the low frequency knee low enough.
The freq points of the 0.025 uf / 1M coupling stages is 6.4 Hz while the 1.5uf / 80k I have in the middle section is 1.32 Hz

I used a 6.8 uf coupling cap to the OPT since I had a nice WIMA MKP to use.
I haven't done any freq sweeps or serious listening yet, so not necessarily official yet.

With the Jensen's, I have 9 dB less in transformer gain than an original but 6 dB more without the filters, hence 3 dB less than an original. As I said above, by reducing the last gain resistor, I eked out a little more gain so my build maxes out at 78 dB measured.

Max gain of original design (copied from another post):
TX:30dB, V1:22dB, V2:33dB, filter:-6dB, V3V4:16dB, TX: -19dB = 76dB max total

This build:
TX:20dB, V1:22dB, V2:33dB, filter:0dB, V3V4:16dB, TX: -18dB = 73dB max total

Note one of the reasons to not do the combined pad/gain switch is it will not work well with a DI. A direct input to the tube grid is one of the best things to have, IMO, but the pad only operates on the balanced input and won't effect a DI.
As I built it, the gain range for the DI is 20 dB lower since it doesn't get the input transformer gain, which works fine (10 dB up to 58 dB)
 
The filters you left out were to restrict the frequency response for TV studio floors so for normal audio studio work they are not necessary, so you did the right thing.

Congratulations on the noise levels you achieved, no mean feat!

One observation on your build.  I have found that toroids are best mounted vertically for lowest hum, but I realise that the one you used was too wide for your case, still the MuMetal shielding did its job.

Thanks for your input.

DaveP
 
Hi
That is a great build, congratulations!
How did you implement the di?
Do you switch in a grid resistor with a relay while disconnecting the input transformer? 
Thanks 
Tobias
 
Back
Top