V76s build - work in progress

[My3gger]

Can't quote properly with this old phone.
Only my HJFP had DI pad, i'm not sure if designer of the pcb included it for running line level through it. G9 is much better example for mic, DI, line switching and powering.
Heater regulation allows slight underheating of tubes, make DC for relays, also eases use of non standard voltages (to certain degree). I know voltage can be drop with resistor, i find it better to regulate if DC is needed. Some cheap heatsinks are made to take two regs allowing running them cooler.
Oliver Archut once wrote about EF804/6 having double helix for less noise coupling with AC powered heaters, new production tubes he inspected at the time didn't have it. Can't say how much difference has such heater in this tubes.
Friend probably didn't use conductive tape or got non heat treated material. Did you ever use it to shield input transformers? I rather leave inputs to professional, although friend will be interested in knowing your experiences with this. Quite a few years ago CJ wrote that bending and cutting ruins shielding properties of mumetal, same can be found in some literature.

 

[Ilya]

Some great progress here.
I've assembled and tested the preamp and all functions are working as expected. I managed to goof up the gain switch connection, but a couple of bodge wires solved the problem (although I'll need to respin the PCBs for the next batch - if it happens, that is).

The only problem that puzzles me is the following. I'm using Sowter chokes that are a bit on the low side in the resistance department. I had put in compensation resistors to make the total resistance of the chokes match the original. The curious thing is the unit works fine with these resistors in place. However, if I short them out, the preamp oscillates at around 230 kHz. As far as I can tell, the oscillation happens in the first stage of the circuit.
What may be the reason for this behavior?

P. S. I promise to post some pics when I dress all wires and make the unit pretty)

 

[Ilya]

Here are some photos.
I’m measuring the performance and it’s truly excellent.
I’ll post some graphs later

 

[thomasdf]

beautiful build ! congrats !

 

[groselicain]

Wow, looks incredible! I've been watching this thread like a hawk and this doesn't disappoint!

 

[Ilya]

The curious thing is the unit works fine with these resistors in place. However, if I short them out, the preamp oscillates at around 230 kHz. As far as I can tell, the oscillation happens in the first stage of the circuit.

Here's the update on this issue.

The oscillation problem was due to the capacitance of the anode choke. It was installed with the end of the winding connected to the anode (as I read somewhere in the threads here, where it was suggested to do so for the lowest capacitance). Well, it turned out that the lowest capacitance point in the choke is the BEGINNING of the winding (doh! obviously). So I should've connected the beginning of the winding to the anode instead. Doing so eliminated the oscillation completely.

Huge thanks to Brian Sowter who took time to respond to my e-mail and pointed me in the right direction.

 

[leadbreath]

Here's the update on this issue.

The oscillation problem was due to the capacitance of the anode choke. It was installed with the end of the winding connected to the anode (as I read somewhere in the threads here, where it was suggested to do so for the lowest capacitance). Well, it turned out that the lowest capacitance point in the choke is the BEGINNING of the winding (doh! obviously). So I should've connected the beginning of the winding to the anode instead. Doing so eliminated the oscillation completely.

Huge thanks to Brian Sowter who took time to respond to my e-mail and pointed me in the right direction.

Hi Ilya,

Is Brian still with Sowter, yhought he left after sale to Carnhill??

 

[Ilya]

Hi!
He seems to be. At least he answers the e-mail with technical questions.
However, their general sales support can be very slow. At one point I had to contact Carnhill directly to push things forward.

 

[leadbreath]

Hi!
He seems to be. At least he answers the e-mail with technical questions.
However, their general sales support can be very slow. At one point I had to contact Carnhill directly to push things forward.

Ahh good, he is a very helpful gentleman..
Glad he is still associated with Sowter.

I have also heard that dealing with Sowter/Carnhill sales now is a pain...
Lets hope they get their shit sorted

 

[Falk]

@Ilya
Very nice build and documentation. Thank you! I am on a V76 Clone myself. As usual I started with a spice simulation of the circuit to get an impression of how things behave.

I will use the regulated power supply with dc filament I built for the U23 clone. Got some input TX from Sowter and some vintage Haufe output TX (10:1) from Ebay. I will most probably use Lundahl chokes.

However I stumbled across a statement by @rock soderstrom
"For the E83F there are no well available alternatives, but still some NOS suppliers for relatively little money. EF83 is a different tube."

I think the E83F is a special quality / long live version of the EF83. At least I found this information here:
http://www.r-type.org/exhib/aaa0475.htm
I have about 60 NOS EF83 Telefunken/Siemens/Valvo which I would not hesitate to use. What do you think?
Regarding the EF804S - trying out reissues (JJ EF806s) in the U23 build I found the best way to substitute this tube is to use EF86. I spent the last Sunday matching tubes. I had 8 new JJ806s from the same batch where (under same conditions) the weakest delivered only 50% of the current of the strongest. Then I had 40 used vintage EF86 where the weakest delivered 75% of the current of the strongest. The maximum emission was about the same. Also I could reduce noise levels tremendously in the U23 build by switching from JJ to vintage Siemens. The U23 has a noise floor of -94dB(FS) RMS and non of the 4 transformers and 2 chokes is shielded.

This being said, I would really be interested in your experience with tubes in the V76 build.
All the best,
Falk

 

[rock soderstrom]

However I stumbled across a statement by @rock soderstrom
"For the E83F there are no well available alternatives, but still some NOS suppliers for relatively little money. EF83 is a different tube."

I think the E83F is a special quality / long live version of the EF83. At least I found this information here:
E83F @ The Valve Museum

Hi Falk, EF83 and E83F are not the same tube. This is a mistake on The Valve Museum website. Pinout, technical data + heater current a different. E83F is one of the few tubes (EF80-E80F as well) where the designation system for the SQ variant does not work. Check the datasheets and you will see that they are totally different tubes.

Edit:

The more I think about it, the more I realize that this SQ designation system only applies (not always) to european ("modern") triodes. SQ pentodes are often designated with three-digit number codes like EF80-EF800, EF86-EF806. However, this is also not universally valid, the European tube code does not have 100% consistency and has numerous exceptions and special cases. This makes it somehow interesting.

 

[Falk]

Check the datasheets and you will see that they are totally different tubes.

I totally agree. Thanks for the clarifiction.

 

[Ilya]

I think the E83F is a special quality / long live version of the EF83

As stated by rock soderstrom E83F is not the same as EF83. I don't think you can interchange them without circuit modification.

As for E806S, I'm using new Tung-Sol tubes. I don't have access to any other versions of 804 or 806 tubes, so I can't compare. Anyway, the total noise is most often dominated by the very first amplification stage, so it's possible to swap tubes in that position and see how it turns out. I'm waiting for the 1:10 input transformer before making any noise measurements though. However, the noise floor is very low.

The U23 has a noise floor of -94dB(FS) RMS and non of the 4 transformers and 2 chokes is shielded.

This really doesn't tell me anything, because I don't know your 0dBFS voltage level.

 

[musicnaudio]

This build looks awesome. Nice job! bet it sounds great.

 

[Ilya]

I've tested three different input transformers and want to share results. Transformers tested were: Sowter 1004, Cinemag CMMI-10C and Cinemag CM-9994A. All transformers were tested for frequency response with their primaries in series (normal Z) and in parallel (high Z). Here we go....

Sowter 1004, normal Z:

Sowter 1004, low Z:


Clearly, given that Sowter is almost 6dB down at 20kHz even with primaries in series, it's NOT suitable for this circuit due to relatively high capacitance.

Cinemag 9994A, normal Z:

Cinemag 9994A, low Z:


9996A behaves well with series primaries. But in low Z connection frequency response falls almost 5dB at 20kHz which can be kinda-sorta usable.

CMMI-10C normal Z:

CMMI-10C low Z:


CMMI-10C is rather close to 9996A frequency wise. But since it has a larger core, it's capable of much higher input levels and lower THD. It's also quite more expensive than CMMI-10C.
I haven't decided which of the Cinemags to use yet, but they both look like a much better fit than Sowter. Yes, the total gain of the V76 is almost 8 dB more with Sowter, but frequency response at high frequencies is not acceptable. It looks like Sowter makes their transformers and doesn't actually tests them in the real circuit which is unfortunate given their price and advertising as a replacement for particular circuits.

I hope this information will be helpful to people who are planning to build their version of the V76.

 

[leadbreath]

Excellent work, this will be very handy...
Bit suprised the Sowter came in with Freq range so crappy, I have used them on Replica V76's and they always sounded good!!

 

[Ilya]

Have you measured the actual frequency response of your V76?

 

[leadbreath]

No, as dont have the facilities at moment.
Need to get something set up, but at the end of the day it will alwsys depend on a listening test, my ears are the deciding factor.... Even though the older iI get the more dubious the results are

 

[dmp]

I would expect those transformers to be flat, not rolling off the highs so much. Maybe a problem in you circuit or test setup? Have you posted the schematic you are using? Did you remove or bypass the frequency filtering components? (Both between stages and between the transformer primary windings)

 

[CJ]

The Jensen 1:10 works very well in the v76. I told Ollie this and he started to offer a 1:10 option on his 76.

You lose some gain but the 76 had so much gain it is crazy.

Sowtet might roll off at 20 k but remember the stock 76 with radio limited bandwidth also rolled off at 20 k.

Sowtet uses a 2 chamber choke bobbin, original was 3. And I doubt they use #48 wire. Thus the lower dcr, but the 76 is so over designed that a drop in choke inductance will probably not be noticed.

If you want to try something crazy, try running the amp with the nfb disconnected and a condenser mic. You will need to put the mic in another room.

Did you mount the sockets o n rubber grommets? First two stages on my amp.