new project, new challenges -> CC IT for 100TH driving GM

Using tertiary windings to offset DC.
Post Reply
JeffreyJ
Posts: 321
Joined: Fri Jan 28, 2005 2:53 am
Location: stuck in the USA for now...
Contact:

new project, new challenges -> CC IT for 100TH driving GM

Post by JeffreyJ »

well, as things go.. project is different now.. the amp with the 10 driver was finished up and shipped out.. I definitely miss it.. I think it was the best amp I have ever built.. globe 50's for output.... beautiful tube, the 50...

so I am now putting together schematics and iron for a GM70 amp... I really wanted to make it two stage without requiring 10Vrms from the driving stage... so the 10 was out.. I also wanted to share the power supply and I hate dropping resistors.. another vote against the 10...

but I like beefy drivers.. you really hear it in the sound.. so I pulled some 100TH's out... mu 38.. but high Rp.. can handle any voltage that we would give the gm70.. so let's pursue...

trying to wind a plate choke.. if I drop the current to 20mA, the lowest I would feel comfortable with at such a high Rp, then the Rp rises to 18k5.. so we are looking at say 800 Henries at 20mA.. big, expensive.. and I'm still not excited about it...

so then the cancel winding bit pops back into my head... and I think this is a beautiful idea.. we can then run much more current through the 100th as we can always just cancel it! the Rp drops back to around 10k.. win, win.. only the power supply suffers a small hit and I like high current in my supplies.. I usually refer to that as regulation... :^)
___________________

ArgH!... just peeked at the curves.. the other reason I was running the 100TH light was to get bias! duh.. it looks like unless we want to run the GM70 higher voltage, that we are stuck around 35mA to bias the 100TH at 10Vgk and 900Vpk... hmm.. that puts Rp at 15k, which I am not happy about, but is better than 18k5... 35mA at 15k is still a bit of drive..

now the only concern is the bifilar style that I like (at least on paper) for IT's..

Dave, what would you recommend here?

GM70 at 800Vpk, 120Vgk, 100mA, so the IT has to handle 85Vrms...
100TH at 900Vpk, 10Vgk, 35mA, Rp of 15k...

of course, I am always open to ideas....

Peace,
Me
nickg
Posts: 29
Joined: Sat Mar 19, 2005 1:46 pm

Post by nickg »

Hi,

I may be missing something, but I think at 800v on the plate, you will find 100ma requires more like -70 on the grid of a GM70.

Or I may be misreading your post, in which case, sorry.
JeffreyJ
Posts: 321
Joined: Fri Jan 28, 2005 2:53 am
Location: stuck in the USA for now...
Contact:

yep, you are right... :oops:

Post by JeffreyJ »

just revise to 70Vgk.. which lowers the available B+ a fraction... I honestly can't recall which tube I was thinking of... I have been scribbling lots of big tubes lately..

maybe that was 845 run HV? 212? 304tl? who knows..

but it has implications! so now I can take one of two routes..

1 - take gm70 higher in voltage and use a 10k OPT and raise the voltage and current for the 100TH.. :^)

or 2 - use a 75tl as a driver.. mu 12 makes the amp 6V sensitive.. which is cool...


hmmm... the big difference becomes POWER... and cap voltage ratings..

thanks...

Me
dave slagle
Posts: 2102
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

but I like beefy drivers.. you really hear it in the sound.. so I pulled some 100TH's out... mu 38.. but high Rp.. can handle any voltage that we would give the gm70.. so let's pursue...
Almost all of the eimac tubes can handle the voltages in question... Maybe you want to half way between the 10 and the 100TH. Something like the 25T or the 35T (3C24?) iirc all have mu's around 20 and GM's of 2500. Which puts you in the 10 ballpark.
so then the cancel winding bit pops back into my head... and I think this is a beautiful idea.. we can then run much more current through the 100th as we can always just cancel it! the Rp drops back to around 10k.. win, win.. only the power supply suffers a small hit and I like high current in my supplies.. I usually refer to that as regulation... :^)
I have been slowly trying to model the cancel circuit in LT spice and it seems a natural to use with any of the eimac tubes since you need a nice DC supply for the fils anyways.
ArgH!... just peeked at the curves.. the other reason I was running the 100TH light was to get bias! duh..
I used the 15E to drive my GM-70 it required 5V to get to full power but I could only get about 600V on the plate, The 25T worked too, but I couldn;t get the plates to cherry with the power supply I had. I used a cancel current choke to load the tube and had 300hy's, I only wish I had thought to use the filament current at the time.
now the only concern is the bifilar style that I like (at least on paper) for IT's..
Well bifilar is out of the question... Unless you want to hipot test them for me :-)
Dave, what would you recommend here?
I hate you guys :-) why oh why must you insist on 1000V tubes and transformers? (of course shines needs to outdo everyone and do 2KV) seriously, I guess we would have to go with a multi section approach.

My gut feel is the 49% nickel core I used for your 10 proto will not cut it, and a bigger nickel core in 49% isn't gonna show up cheaply. So maybe we need to switch the design to amorphous. The core I used for the 15E plate choke would work well. Plus it allows for a three chamber bobbin so the cancel winding has its own bay...

I think using steve's idea (see his D3A EL34 amp driver) of a tertiary winding in the cathode to bias the tube deserves a good looking into also... I think there is a model for the SV572-30 around, it has double the Gm of the eimac stuff, but should give us a picture of what is possible.

JD should also have some input on a good tube choice, maybe pick his brain for a good choice... Seems the two of you could help each other out with R&D for the iron and then we will have independent results for the same idea.

I also agree with nic... I found the GM-70 to be a "double 50" where the 50 biased up at 50V 50ma and 350V, the GM70 liked the number 100. 100V 100ma and 100W of Pdiss (from a 100 decavolt supply)

dave
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

I think there is a model for the SV572-30 around, it has double the Gm of the eimac stuff, but should give us a picture of what is possible.
Just a quick note... For the LTSpice stuff, if you use that EMISSIONRANGE parameter I inserted, you can pretty much get exactly what you want above. For example, using the SV572-30 and setting EMISSIONRANGE to 0.5 produces a "tube" with exactly half the gm of the "stock" 572-30, but exactly the same mu. I think that's what you are looking for, unless I'm reading something wrong.

Thus, using EMISSIONRANGE, you can check all different kinds of tubes, from modeling ones that otherwise have no model defined to coming up with your own fictional tube. :D

-Steve
JeffreyJ
Posts: 321
Joined: Fri Jan 28, 2005 2:53 am
Location: stuck in the USA for now...
Contact:

revison and open to thoughts..

Post by JeffreyJ »

ok.. the 100TH is out.. so I am thinking 75tl or VT127A (which we think is a 100TL with a 250 filament)..

the GM70 looks pretty good at 900V and 100mA...

sorry about the high voltage.. honestly, Dave.. I am usually your small tube ally.. I am tired of people telling me I need to build a "real" amp.. so I am.. if it beats a TJ 2A3 mesh or a vintage globe 50, I will cry uncle.. and build big.. if not, small again for a while..

I will polish my circuit up a bit.. I really wish we could bifilar.. no hipot testing for me.. maybe since it is 1:1, a bridging cap?

Peace,
Me
dave slagle
Posts: 2102
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

ok.. the 100TH is out.. so I am thinking 75tl or VT127A (which we think is a 100TL with a 250 filament)..
Cool... Makes the IT a bit easier.

I am tired of people telling me I need to build a "real" amp.. so I am.. if it beats a TJ 2A3 mesh or a vintage globe 50, I will cry uncle.. and build big.. if not, small again for a while..
I built big, then I built small, now I'm itching to get big to sound as refined as small.
I will polish my circuit up a bit.. I really wish we could bifilar.. no hipot testing for me.. maybe since it is 1:1, a bridging cap?
Well I consider the bridging cap a cap coupled circuit... If you are game for that so am I... Lets see where the circuit leads us, then we can try a few things.

dave
JoshK
Posts: 13
Joined: Fri Dec 02, 2005 4:37 pm
Location: NJ/NYC

Post by JoshK »

What ever became of this alley? Curious minds want to know.

I am brainstorming/sketching mode for a 3C24 > GM70. Was looking at what has been done. Did you guys ever build such an CC IT? How did it work?
----
Josh
JeffreyJ
Posts: 321
Joined: Fri Jan 28, 2005 2:53 am
Location: stuck in the USA for now...
Contact:

Post by JeffreyJ »

dave wound it.. I built it... I ended up using the Eimac 15e to drive the 304tl... the driver stage was amazing.. super clean... I am actually rebuilding the power supply side of the 15E breadboard so that I can try the 15E in my phono stage...

in the end, the combo of Eimac output and driver was a little too "white" for my tastes... an oxide driver ended up sounding much better... more relaxed... less hyper... but I kept the nickel 1:1...

I highly recommend offset winding equipped, bifilar wound 1:1's from dave... just try to keep the Rp of the tube below 10k if you can...

Peace,
Me
JoshK
Posts: 13
Joined: Fri Dec 02, 2005 4:37 pm
Location: NJ/NYC

Post by JoshK »

Thanks for the reply. I am definitely willing to try when I get there. Its convenient that Dave is nearby.
----
Josh
JoshK
Posts: 13
Joined: Fri Dec 02, 2005 4:37 pm
Location: NJ/NYC

Post by JoshK »

I am curious how you determine operating point when data is so sparce. Do you just determine it in situ? Throw some volts on the anode, play with the bias and watch the anode current?
----
Josh
JeffreyJ
Posts: 321
Joined: Fri Jan 28, 2005 2:53 am
Location: stuck in the USA for now...
Contact:

Post by JeffreyJ »

well... first build the filament supply... a variac will help here.. build for 6V and variac down...

then build a high voltage supply... again use a variac...

clip leads and a stack of cathode resistors will get you where you want to be...

clear as mud?

Me
DowdyLama
Posts: 104
Joined: Thu Jan 27, 2005 5:42 pm
Location: Blue Ridge Mtns, NC
Contact:

Post by DowdyLama »

Or, you can do like I do: use an existing DC filament supply/variac, and...this is the important part: decide what a *convenient* B+ would be and start with that 8)
Then clip in a 1k cathode resistor and see which direction you need to go in re: bias.
dave slagle
Posts: 2102
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey Josh,

Jim and JJ have it covered thus far. I will try to steer you away from the tube you want to try for the 1:1 offset IT because the bifilar nature and the 500+V supply are at odds with each other. I like to keep the B+ on the bifilar pair at 200V or below and will condone slightly above that if I have the proper wire on hand.

dave
JeffreyJ
Posts: 321
Joined: Fri Jan 28, 2005 2:53 am
Location: stuck in the USA for now...
Contact:

Post by JeffreyJ »

DOH!

yeah... I forgot about the bifilar voltage issue... I have an arrangement with dave.. if I break, burn, explode anything, it is my fault.. and I have to report the fault conditions.. I guess I am more of a beta tester on some things...

still.. I'd aim for 350V and a 3c24... then make him be really careful when he winds...

dave, feel free to punch me later...

Me
JoshK
Posts: 13
Joined: Fri Dec 02, 2005 4:37 pm
Location: NJ/NYC

Post by JoshK »

Thanks guys. Basically what I thought re: biasing.

I saw the point about bifilar before. What makes the bifilar sound better than seperate wells? (I am pretty newb when it comes to magnetics)
----
Josh
nic
Posts: 12
Joined: Thu Jan 19, 2006 8:59 am
Location: London UK

Re: revison and open to thoughts..

Post by nic »

JeffreyJ wrote:ok.. the 100TH is out.. so I am thinking 75tl or VT127A (which we think is a 100TL with a 250 filament)..
Ran them zero bias at 600V , biased up to around 75mA if I remember correctly . Thats more like a 75TH than a 100TL . IMO the VT127A has a 250TL filament with the specs of a 75TH with 100W dissipation . Watching this thread with interest :)

cheers

Nic
dave slagle
Posts: 2102
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

JoshK wrote: I saw the point about bifilar before. What makes the bifilar sound better than seperate wells? (I am pretty newb when it comes to magnetics)
Well it just does :-) If you want me to hazard a guess as to why, I'd speculate that the wide bandwidth nature of the bifilar nets you a marked improvement. 100Khz+ bandwidth without peaking or the need for loading is par for the course with the bifilars. I have found that transformers that peak near the audio band always seem to sound harsh, and the only thing that sounds worse is a transformer with a load applied to tame the ring.

dave
Post Reply