Hypothetical Attempt at V76M boards

[Potato Cakes]

Hello, everyone,

I have been thinking about making boards for a stereo V76M preamp for sometime but have yet to give it any serious thought until recently. The idea is to do a layout with both preamps and the PSU (not the transformer) on the same board. Here is the schematic I am using:

http://audio.kubarth.com/rundfunk/getfile.cgi?f=A%2C38Q%2CS8P%2CC%3CT-%25%5D%22%3CF%25U%3BF%29U8V%40O%3D%26%25B7U8W-FTN%3C%261F%0A

The 2x12 gain switch would stay as it is. I would do away with the fine gain switch (0 to +5dB) and replace that with a switch that just does +3dB. Also I would do away with the 1dB gain control as well. If you need precision gain control you would just do that in a DAW. I have been reading over on an old V76 build thread here and noted that the choke (BV Nr 517) can be omitted as it affects HF response. A DI section would be nice as well as some way to do a simple variable HPF, but I am fine going without one.

I am curious about the capacitors labeled 35 and 37 being electrolytic. Is there a reason that these would not be film besides the fact that when this circuit was new that a 2uF and 4uF 300V film cap would have been too big? Also I am curious why there is a trimmer for the heater voltage just for V1 and separate transformer winding for V2-4. Does this also have to do with the components that were available at the time? I also saw on CJ's V76 build that some of the inductors were replaced by resistors. I am wondering if that could apply here, especially for BVNr 515. Lastly, I would want to have the transformers mounted on the board. I am considering the Carnhill 9045M and/or Lundahl 1578 for the input and haven't decided on an output option. Thoughts?

Thanks!

Paul

 

[Potato Cakes]

Hey Paul

No reason to not change electrolytic capacitors to film, size and price notwithstanding.  There's more chance the value will be closer to what's written on the tin, and they're unlikely to ever need replacing.

If you use quiet regulated DC for the heaters then one winding for all valves is fine. 

Replacing the plate load inductors with resistors would work, but you'd need a higher B+ to feed those resistors to maintain the same operating point. 
Replacing BV515 especially would require another 200 volts give or take, otherwise headroom would be diminished.  All the data you need for the operating point is on the schematic, almost 12mA on the E83F anode and 170V DC. 
An inductor, besides the stored energy of swinging twice the voltage, also shifts the load line.  A constant current source does too, in a similar way.  Except a pentode's anode is already a CCS, so then you have 2 CCS's possibly fighting each other so you need think about adding a shunt resistor. 
None of this is insurmountable and you may end up with a better performing/testing unit.
Just food for thought.

On choices for transformers and chokes, I used all Cinemag on the Mercury M76m.  But I bet there are lots of good choices out there. 
I like the Lundahl anode chokes as you can specify DC current vs signal voltage vs inductance, but they can be pricey. 

Thanks for the info, Winston! Perhaps I will leave the BV515 and BV514 in there for now as I would have more investigating to do to figure out the math and how to implement it. I'm fine with formulas but I have difficulty figuring out which values from the schematic to use if they are not clear for someone like myself. I do recall from that thread I mentioned that BV517 was to be omitted as it would affect the HF response. It also does not seem to be used as a plate load. It is also feeding the grid of V3 and the tube preamp schematics I have studied do not have an inductor in series with the input path.

DC regulated heaters will definitely be used.

I'll keep looking at transformer options. I'd like to do a footprint for 2 or three options for the input transformer and just go with a quality and probably end up wiring the output transformer off board. Right now I'm not worried about layout as much as I am getting the schematic correct. I'll post my proposed edits in the coming week. I hope.

Thanks!

Paul

 

[Winston OBoogie]

Here's a link which shows you the relationship between DC current, AC signal, and inductance.  It pertains to the Lundahl LL1668, but similar relationship rules apply to any plate choke.

Anyway, good luck dude

https://jacmusic.com/lundahl/applications/How-to-bias-the-LL1667-LL1668.pdf 

 

[Potato Cakes]

Here's a link which shows you the relationship between DC current, AC signal, and inductance.  It pertains to the Lundahl LL1668, but similar relationship rules apply to any plate choke.

Anyway, good luck dude

https://jacmusic.com/lundahl/applications/How-to-bias-the-LL1667-LL1668.pdf

Thanks for the link. I will give this a thorough read.

The S version schematic I have is a bit illegible but I do see the send/return lines that you mentioned. If I was going to add a variable HPF in front of V3, I'm not quite sure how to deal with the insertion loss. I noticed there is a different value of the resistor that connects V2 grid 1 and ground for V76 (1M) and V76m (500k). Is there where the V76 is accounting for the HPF that comes before V3?

I really appreciate the help.

Thanks!

Paul

 

[Ricardus]

The schematic won't load for me.

 

[Potato Cakes]

The schematic won't load for me.

http://audio.kubarth.com/

You'll have to go through the menus. Apparently the direct link is only good once.

Thanks!

Paul

 

[Ricardus]

Thanks!

Well, I think many people would be interested in this project.

 

[shabtek]

CJ?

 

[Potato Cakes]

Here is my redraw for the schematic to use a fader per Winston's suggestion. I added the 15k to the 50k resistor (64 and 60 from the original schematic) so that I could get -infinity to +5dB when fully CW, however I don't know is if the value of the pot will be high enough. I also took into account the resistor and potentiometer from the 1dB adjust and added that to the 80k (75) and then rounded to 100k for simplicity and ease of sourcing parts, but I don't know if that needs to be changed at all.

I also noted that on the V76 schematic, the grid 1 to ground resistor (61) is 1M which also has a passive HPF/LPF section and the M version has this resistor value as 500k, which doesn't have the filter section. My thought is if one wanted to add a HPF then that resistor would have to be changed to 1M per the V76 schematic.

For the gain switch, I'm wondering if I should make use of that last position and either do 65dB or -6dB. Is either of those possible by adding a resistor in either direction?

Thoughts? Corrections?

Thanks!

Paul

 

[Potato Cakes]

Original V76m for reference.

 

[Potato Cakes]

Yep looks good for the fader spot Paul.  The 100K fader along with what you now show as a 65K series R forms a divider &, as is,  gives you a bit more than an extra 6dB gain at the fader top position so you're at the 66db total + a bit. 
Adjust the value of your 65K once you have a prototype done to dial it in.  I think I ended up with 80K but it's been about 15 years so...

I would also include a spot for where the 15K R is on the "M" schematic (part 64) as I seem to remember it being necessary.  Might be that 1K is enough and you can locate it along with the 2uF cap close to V3's grid on your pcb. 

I personally would just use 1M for part no. 61.  There are enough low end poles and enough low end phase deviations already happening in the circuit that, this one being an octave lower won't hurt, can only help.   

On part 75, originally 80K, now 100K on your schem, the difference in gain between the two is less than 2dB so not anything to worry about that much.  Again, wing it for now and adjust if necessary when you have a proto

Biggest issue is finding good chokes at affordable prices really.  I checked the current prices on Lundahl and it's about 123 euro whatever that is in good ol' US dollars. 

Sowter have ones that are supposed to be good (haven't tried them myself) but their situation is in flux (non intentional pun, sorry  )  with being bought by Carnhill. 

My vote is to just leave the switch at 11 positions for what little that's worth.  I'm actually not convinced the lower level positions which incorporate the input pad are best with the "M" schematic values.  The "S" and other V76's used values which presented a more constant lower impedance to the input transformer. 
Dunno, but check out the S and the other versions.

I personally used the values from the EMI REDD pre-set attenuator for my levels below 34dB gain, but it probably made zero practical difference to the way it was done on the V76S tbh

Winston, this is really great information. Thank you!

The thought with removing the 15k and adding it to 50k was that when the fader was fully CCW then it would make a short the signal going to g1 of V3 to ground, completely silencing the signal i.e. negative infinity. I didn't know if it still would do that with the 15k where it was but I will definitely leave a spot for it.

I did see Simon's build. It's very impressive.

I'm going to do some more reading and look at replacing the chokes with resistors/caps and see if I can get around needing inductors. If I can achieve very similar results and save a noticeable amount of money on building these then I think it will move this project along a little quicker.

Thanks!

Paul

 

[Potato Cakes]

I personally would just use 1M for part no. 61.  There are enough low end poles and enough low end phase deviations already happening in the circuit that, this one being an octave lower won't hurt, can only help.   

You mentioned this above. I read it and immediately forgot it!

Thanks!

Paul

 

[shabtek]

wonder if PRR has just been busy moderating the IRS forum...

 

[Potato Cakes]

P.S.  This is right up PRR's alley, he can do this stuff while also filling out everyone else's tax returns.

I'm less of a "doing all the maths beforehand" guy as that stuff was never my strong suit.
I'm quicker just lashing it up and adjusting stuff until it fits.
If I get stuck then that's when I take a time out and do all the sums.

There are other dudes on here who excel at that stuff though so, hopefully someone comes along to offer a hand.

I usually have to resort to strapping everything together then figure why it didn't work as I don't have a reference of the various formulas needed and when they are applicable.

I think the best solution to move forward with getting an actual board laid out is to leave a spot for a resistor or an off board inductor as those would be sizable enough to make sense to keep it off the PCB. This is ultimately a DIY project so there will be room for additions and substitutes per the builder's taste and imagination.

Also, following further review I think it would be better to keep these as single channel boards with an off board PSU. This will make this circuit accessible to more people who may only want a single channel.

I still have some reading to do regardless...

Thanks!

Paul

 

[Potato Cakes]

Yeah.  I have seen him posting on DIYAudio recently so it seems he's doing OK which is good to know.

Paul, I think your idea of placeholders for resistor or off board inductors is good.

I always include the variables in a first board run myself - better/easier chance of it working without having to hack it.  Jumpers or omitting a component is easier than cutting traces.

You can always do revisions on a 2nd run if needed.  I bet there'd be enough folks who'd take any initial boards left over from a 1st run so you can cover your cost.  As long as it's got pads for options/getting out of jail etc.   

Like you say, this is DIY

Just a show of hands, how many people would be interested in something like this? If I can cover costs then I would be more motivated to get this done sooner.

I'm going to contact the guy that has been doing PCB design for me in the past and start working on an estimate of initial expenses.

Quick tangent, will running the tubes with 12VDC heaters cause a significant shortening of the tube's life or performance? I'm trying to sort out the PSU requirements and the spec sheets for EF804 and 83S rate their current draw based off of 6.3V but there isn't a range of nominal voltages for the filament. If increasing the voltage will have less draw on the PSU then I would prefer going that route if the penalty for doing so is negligible. I don't want to derail myself with another topic and if the short answer is "don't do it" then I'll leave it at 6.3V and press onward.

Thanks!

Paul

 

[Ricardus]

I am interested but can't help with $$ for about a month when my tax return comes.

 

[Potato Cakes]

I heard back from Lundahl regarding chokes and they said they do not to custom windings as they do not offer inductors with values that high. I am still waiting to hear back from Cinemag.

Thanks!

Paul

 

[Potato Cakes]

Lundahl sent a follow up email and it seems that their 1667 will do 250H and 400H at the current draw amounts per the schematic. Sowter said to expect a 12 week lead time for their 1377 and 1378 chokes but possibly sooner as they are taking on more staff.

No word from Cinemag yet.

Thanks!

Paul

 

[Potato Cakes]

Thought about this last night, did a few load lines, some reading up, maths etc and -  I believe both of the valves in the V76 that are choke loaded are basically acting more or less like triodes that have a CCS in the anode.  Strapping both valves as triodes and putting some solid-state stuff above them should get them acting like they are with the inductors.
A resistor would get you a bit of the way there, fed from a clean 350V H.T. supply, still with the valves strapped as triodes.  But those inductors help linearize the 2 valves and also raise the gain close to the triode'ed mu of the valve.  It'd be a bit sloppy with just resistors.

Needs an active load also fed from that same 350V B+  .  Tube or solid state active ccs, either would work

Edit:  Or -  Make each stage out of cascoded pairs and plot the load lines individually for top and bottom.  Then apply ultra-linear feedback to convert to mostly triode.  The E83F isn't quite all triode but that 2nd EF804S is about spot on. 

Can't get through all the maths to figure out what the supply might need to be then, but gut tells me it'd probably have to be higher still...  like 470V DC.  Ouch!

Anyway... 


Edit

470VDC is definitely quite high. For myself I think I'm going to stick with inductors at this point. The pricing I got from the US distributor for Lundahl is very reasonable. I still would be interested in exploring the use of plate resistors but maybe that could be a followup thread after I hopefully get some boards made sooner than later.

I'm working on a rough layout that I'll post later this week.

Thanks!

Paul

 

[rock soderstrom]

Lots of good Information in this thread. Thanks a lot