Ultralinear for Single Ended amp?

[jordan86]

 In my never ending curiousity with my little single ended amp, it just dawned on me that the idea of using an ultralinear OT has never crossed my mind. Had the thought today and now I’m wondering if anyone has done it, or what could be expected with that “upgrade” in a SE scenario??? Very limited results on the search function. Probably for good reason :)

The amp is a tweaked AX84 project. Started off as their SEL (sorta JCM pre into KT88) for those who are familiar.  Now it’s like a JTM/Superlead preamp into a single KT66. I have the custom 20W SE Doberman iron that the AX84 guys have custom made by Heyboer.

Looks like Hammond has a line of 30W SE UL output xformers. They are not cheap though. Maybe not worth the trouble.

[PharmRock]

Hi Jordan,
I'm not sure if this fits into the category of "ultralinear", but in one of my heavily modded valve juniors (single-ended EL84) I installed a conjunctive (corrective) filter similar to what Dr. Z used in his Carmen Ghia (and also found in the Watkins Dominator)...there were quite a few threads/posts on this on the sewatt forum.  It really smoothed out and evened the response across all frequencies.  I had it switchable and the difference was very significant. 

Again, this suggestion may not be exactly what you are looking for, but it may get you in the neighborhood.  Plus, its very cheap.  Certainly cheaper than a new OT!

Scroll down to post #16 on this thread for sound clips (not mine, but representative of what I experienced). https://www.tdpri.com/threads/the-secret-of-dr-zs-amp-mojo-rca-conjunctive-filtering-design-guide.310494/

hope this is useful

[SILVERGUN]

good read
ultralinear.pdf (tubebooks.org)

[shooter]

I've built 3 with UL OT's, 2 KT88 SE and one EL34 PP - self biased.
the guitarist's liked them well.

[tubeswell]

good read
ultralinear.pdf (tubebooks.org)

For $0.25? Obviously can't put a price on knowledge.


(Edit: skip to the grid characteristics curves in Fig 3a, 3b and 3c on pages 11 and 12 for a helpful intuitive understanding about how UL (ideally) linearises the curves in order to minimise THD)

[PRR]

> For $0.25?

I think Keroes and Hafler were taking money on the transformers, kits, and amplifiers. The two bits was just because people don't value free information.

Ultralinear makes a Pentode more like a Triode. Is that what we want in Rock-n-Roll?

[jordan86]

Hi Jordan,
I installed a conjunctive (corrective) filter similar to what Dr. Z used in his Carmen Ghia (and also found in the Watkins Dominator)...there were quite a few threads/posts on this on the sewatt forum.  It really smoothed out and evened the response across all frequencies.  I had it switchable and the difference was very significant. 

hope this is useful

Thanks PharmRock. I am familiar with the conjunctive filter. I looked it up a while back but stalled at figuring out what values to use, and then moved cross country and so it went on the back burner. That link is very helpful. Thanks.

[SILVERGUN]

Jordan, without knowing WHY you would want to go to ultra linear it's hard to say whether it would be a good idea or not.
What is the design goal?


The word "ultralinear" certainly sounds cool, but I'm not sure it sounds cool. 

[jordan86]

The word "ultralinear" certainly sounds cool, but I'm not sure it sounds cool. 

I’d love to smooth out some of the harshness in this amp and give it some more fullness. Being single ended, it can be a bit raspy and unpolished. Doesn’t quite have the guts I am looking for. But that’s to be expected with an SE to some degree.

Due to the nature of the circuit, adding NFB isn’t an option either. I figured UL might be a good way of getting a more refined tone, but I think that conjunctive filter might be my first route I take. 10X cheaper.

[SILVERGUN]

One other cheap thing to try would be to increase the screen resistor with the intended result being more compression.
If your at 1k (guess) you could reasonably double that and tell us if you can hear a difference.

[jordan86]

One other cheap thing to try would be to increase the screen resistor with the intended result being more compression.
If your at 1k (guess) you could reasonably double that and tell us if you can hear a difference.

Here's the schematic.

[SILVERGUN]

Here's the schematic.

I'm not sure how you got there but something in the preamp looks out of place to me.
You might want to clean that up for us old guys with old eyes.

I see it now

[jordan86]

I know it's a mess. But basically, it's just the first two triodes in parallel. I don't have a schematic building software. I took the two first stages that were in series, made them parallel, and tweaked the voltage dividers into mixing resistors. So went from a JCM800 style pre to more of a JTM45 pre. It looks hideous on paper though. :)

[jordan86]

So if I do the conjunctive filter, it goes from the plate of the output to B+1? May gator clip in to test. If I like it, I'll try putting them on an SPST just to be able A/B. Attached an old picture of the guts, pre-mod days, of where I may put it. I know the schema says they are brown and blue, but I have an upgraded OT so its Brown / Brown-Yellow. Toggle would be up through the top side of the chassis.

[shooter]

Can't recall, do you have a scope?


IF so;
look at what you're feeding the PA stage
play with the knobs and watch again


sometimes "that raspy unpolished" sound is ugly stuff from the Pre.


I try and feed the PA with a clean, (TS adjustable), signal equal in amplitude to the PA tube's Cathode DC volts, so;
if Vk = 32vdc
I make sure there's a clean-ish AC signal around 32Vac-rms feeding the 88
I try and get this signal with gain n MV pots set at about 7
that gives the guitarist some room to push it into the Un-polished range 

[jordan86]

I don’t have a scope, sadly. I had access to one when I was in Colorado and a friend who was good with it, but not anymore. I could be wrong but I don’t think it’s is anything too ugly or erroneous. I may just be asking more from a SE amp than is fair to ask. It does sound good through my Creamback speakers. Just looking for those last little bits of “polish”. I like the idea of the conjunctive filter. I read up on those and the Zoebel networks a while back. I’ll implement that first and reevaluate.

[SILVERGUN]

I don't have a schematic building software.

I don't either. I usually wind up just using Paint (built into Windows) to cut and paste schem. snips and sections together.

Like this:

[shooter]

I’ll implement that first and reevaluate.

If that didn't get you close enough;
try dropping the 100uF (C6) on the 88 to 25uF - 47uF
quick n easy.  I found 100uF is good sometimes but 47uF is best most times

[trobbins]

In my never ending curiousity with my little single ended amp, it just dawned on me that the idea of using an ultralinear OT has never crossed my mind. Had the thought today and now I’m wondering if anyone has done it, or what could be expected with that “upgrade” in a SE scenario???

Afaik, the earliest known foray in to that form of SE design was back in circa 1932.  They didn't call it UL back then, rather it was called a 'half-triode, half-pentode thing', but it obviously sounded better!
https://www.dalmura.com.au/static/Pages%20from%20March%201955-2.pdf

[tubeswell]

https://www.dalmura.com.au/static/Pages%20from%20March%201955-2.pdf

Here I go off-topic, but that's an interesting PI on Page 4

[PRR]

> that's an interesting PI on Page 4

Half-buffered cathodyne.

What happens when "Grid 2" goes to grid current? (Something like what he's talking about with capacitance, only wide-band.)

[jordan86]

I found some parts in the ol drawer for a conjunctive filter. They may be under spec though. .001uf (630v) and 10k metal oxide (2 watts). Would it be too crazy to just clip those in for some quick tonal experimenting? If I like it, I’ll order some 1kV caps and 5w resistors.

I currently have a KT66 in there. Running at like 63ma and 400v, as best as I recall.

If I smoke a 2W resistor, does it short across/over the resistor or does it break and open the circuit path?   

[shooter]

for every "rule" there's an exception, so;


Resistors tend to burn open most other components tend to burn short, then open if poorly designed safety features fail


tack away, let your nose be your guide

[jordan86]

Well I got smarter...I had multiple 10K/2W metal oxides so I paralleled two. Experimented for a bit with different caps with the 5K and 10K respectively.

For any who care to know...I think .002 and 5K is a good combination. I liked .001 as well, but still just a little bright still. Mind you this is all through a Creamback M, which is a very warm speaker.  I even tried a .001 and a 500pf cap in parallel (which may have been my favorite) but sourcing a 1kv cap at .0015 may be tricky.

Overall, the conjunctive filter is nice. Tames some highs, and adds a sort of compression that feels a little more polished but also evens out the string to string response as well.

[PharmRock]

Well I got smarter...I had multiple 10K/2W metal oxides so I paralleled two. Experimented for a bit with different caps with the 5K and 10K respectively.

For any who care to know...I think .002 and 5K is a good combination. I liked .001 as well, but still just a little bright still. Mind you this is all through a Creamback M, which is a very warm speaker.  I even tried a .001 and a 500pf cap in parallel (which may have been my favorite) but sourcing a 1kv cap at .0015 may be tricky.

Overall, the conjunctive filter is nice. Tames some highs, and adds a sort of compression that feels a little more polished but also evens out the string to string response as well.

Thanks for posting the results...this is pretty much my experience as well.  It tamed the amp a little, made it a little smoother.  The trade-off is losing a little bit of the "rawness" from the amp, but I really liked it for overdriven leads.  I noticed with larger cap values (0.oo47uF) it started to have a "damp blanket" effect on the tone. 

[jordan86]

Yes, .002 would be the biggest I would go, just for the record. I tried a .0068 and it was very muddy.

[PharmRock]

is 630v rating on the cap not high enough for the plate voltage you have?   is the goal at least 2x the plate voltage?

Alao, could you tell much difference between the 5k and 10k resistor in the circuit, or were the changes mainly due to cap value?

[jordan86]

My plate is at 390v, so probably need a 1kV cap. No smoke and fireworks though during testing so that’s good.

Cap seems to make the most difference tonally. The resistor had more affect on the feel to me. 10k felt more compressed and “smeary”. Like there was quite a bit more latency or sag on the attack. Seemed kinda cool at first, but going back to the 5K, I immediately noticed the difference. Had more of that plexi-like punch, which is what I’m going for. Think I will pass on the ultralinear OT for now. Maybe I’ll try it one day if I can find one cheap. I don’t need more power per se. The amp is plenty loud with the master on 5. Wanted to just refine some of the rawness just a bit and this seems to have done it.

[shooter]

I will pass on the ultralinear OT for now.

I don’t need more power per se.

UL will be less power It "falls" somewhere between pentode (more power) and triode (less power)
what i've found, it also "cleans up the rawness" to a certain degree, once you start compressing, or clipping the signal all beats are off.
IF you go down that hole, get a UL OT that is "bigger" than required, can handle more watts than you put out, UL is not meant for OT saturation, it's better at just passing to the speaker what it gets on it's input without adding or subtracting from the signal.

[tubeswell]

630v rating on the cap not high enough

Correct, For a conjunctive filter (which sits across the OT primary), for a ‘belts and braces’ rating, you want a cap voltage rating at least 2x as high as the voltage across the OT primary (which, on an SE OT, is between the output tube plate and the Filter Cap end of the primary winding).


You can get away with less than ‘2x’ in real life because the output tube plate will never be able to swing all the way down to 0V (or all the way up to 2x), but you’d be taking a bit of a gamble

[jordan86]

I’ll implement that first and reevaluate.

If that didn't get you close enough;
try dropping the 100uF (C6) on the 88 to 25uF - 47uF
quick n easy.  I found 100uF is good sometimes but 47uF is best most times

Just tried this and right you are, Shooter. I compared 25uf, 50uf and the stock 100uf.  50uf wins!!! Cleans up the muck without thinning things out too much. Feels more polished and clear. I actually really like how polished 25uf sounds, its just a tad too thin with the Strat. Fun experiment. Thanks for the suggestion.  This and the conjunctive filter have made a major difference.

[madnis]

https://bmamps.com/ivds.html


I've used the above with great sucess.


I've used the schematic your using
(with some variation based on OP tube,
same OT from hammond).


I changed the
100ohm grid stopper to 4.7k. This rounded
the signal off top and bottom. Was good, but not great.
 
I changed the CF 100k cathode resistor to 68k then 82
then 91k. I settled on 82k. This is a nice mod to try.