Transformer selection for a SE KT66 JTM45-esque build

[goldstache]

I want to build up a Single Ended design with a KT66.  I would like to have the dual inputs of a JTM45. 

My concerns:

Finding the right iron
I am not sure which PT to get and what 5v rectifier tube to use for JTM level voltages.  I was thinking 400-430VDC target B+ using a 330-0-330 @125ma transformer I have laying around.  It has a 5v tap at 3A.
I built a SE KT88 amp that used the same transformer and it gave me around those B+ voltages with a similar topology.  I know the load will be different but not drastically me thinks.
At those voltages I calculated the OT primary at 5.6K  (410v X 410v/30W)
Would like to keep the iron on the cheaper side of life.

The lack of PI
I will not have a PI due to SE power amp, so I'm wondering if I will need another stage to deliver JTM45-esque signal voltage to the KT66?

Amp Wattage
I know that a KT66 is a 30W tube in PP configuration.  Not able to find info for single ended operation.  I'm thinking its around 1/2 that??????
I'm trying to maximize output power, which brings me back to the Iron issues.  Still learning how to pick the OT for maximum RMS.

I've looked around but have not found a dead wringer schematic that suits what I'm after.  I've seen some SE plexi type circuits and that's a good reference, but none included XFMR info or BOM's. 

Any help is greatly appreciated. 

Thanks

[PRR]

KT66 is practically a 6L6GC. Use the same values, you will be fine. (Note higher heater current.)

> Not able to find info for single ended operation.

GEC KT66 data (6MB PDF), page 2.

> I'm thinking its around 1/2 that??????

Never that much due to distortion and efficiency difference SE and P-P. GEC claimed 6W as SE but 30W-50W as P-P.

You need at least two voltage-gain stages from guitar to power tube. Tone-knobs or reverb may need 3 stages. You need to sketch-up your thoughts.

[HotBluePlates]

I want to build up a Single Ended design with a KT66.  ...
I'm trying to maximize output power ...

You are not trying to maximize power output, or you wouldn't choose single-ended operation.

How much power do you need?

[John]

I want to build up a Single Ended design with a KT66.  ...
I'm trying to maximize output power ...

You are not trying to maximize power output, or you wouldn't choose single-ended operation.

How much power do you need?

Agreed. I like SE myself, and usually use a 5881 instead of 6V6. For tone, not output. Just my 2 pennies.

[goldstache]

You are not trying to maximize power output, or you wouldn't choose single-ended operation.

How much power do you need?

I just want to get as much out of a single ended KT66 as possible.  The data sheet that PRR supplied has been most helpful.  I am still drawing out the schematic and trying to approximate voltages.  I will have a working drawing soon. 

I have a bunch of PP KT66 stuff, I just have never made a scratch design with 1 single ended.  Thought it'd be a good learning experience. 

I'll post it when I get my act together. Thanks. 

{EDIT - untangled quote -PRR}

[John]

have a bunch of PP KT66 stuff, I just have never made a scratch design with 1 single ended.  Thought it'd be a good learning experience. 

Always the best reason!

[goldstache]

Well I finally got some time to wire this up and check some voltages.  It's breadboarded so I can tweak it.

Here is my schematic as wired.
(One Caveat: I'm waiting for the IRF820's to come in the mail, so I have a stock 12ax7 JTM45/Marshall CF as of now)


The preamp isn't able to drive the KT66 without an additional stage after the tonestack.  So I have a pot after the coupling cap, headed into the grid of the low gain half of the ECC832.

I was really just trying to get the voltages where they are within dissipation before I start working it a bit. 

KT66 Bias math:  here goes nothing.......
Cathode resistor: 470R
Cathode resistor Vdrop: 27.8v
V/R= 59ma (HIGH)

Plate Volts minus cathode voltage= 338.2v
I from above: 59ma
V X I= 20W dissipation

Now for a KT66 the data sheet says it's a 25W tube in pentode.  But for SE it states 7W Pout.  I assume this is just Output power not dissipation.  Hope that's right. 

My PT is rated at only 75ma.  Got any bias pointers for the power section as I'd like to start there before I turn toward the preamp adjustments. 
Any insight is highly welcome

Thanks.
 
 

At first play, it works, but it doesn't sound "right". 

[PRR]

> ...data sheet says it's a 25W tube in pentode. 
> But for SE it states 7W Pout. 
> I assume this is just Output power not dissipation

"Pout" is not just an unhappy face.

In this context it is Output Power.

[sluckey]

You must have a resistor from the output tube control grid to ground.

EDIT... Added... Your schematic shows Q1 gate connected directly to source. That's not right. I know you have a 12AX7 in there for now, but is the grid tied to the cathode???

[shooter]

But for SE it states 7W Pout

If you have a scope or good meter, measure Vac rms at speaker say of a 1khz sinewave, do the math and you'll get actual audio power out, in SE I typically get about 30% of Plate dissipation clean, maybe as much as 40% full bore.

[goldstache]

Good eye Sluckey!  I have a 220K to ground at the KT66 grid.  I did forget it in the schematic though.  Thanks.  I will edit when I get the chance. 


The IRF820 CF design was a lift from you.  Thanks.  I have a 12v zener for the Source to Base.  I should've checked more clearly before posting as It's silly for me to expect everyone to know what is bouncing around in my brain versus what's on paper.  Currently it's just the standard Marshall follower there.  Direct plate, direct grid, cathode 100k to ground.


I always first try and get the power stage biased in and then I work from there. 


So, to clarify, when I calculate the dissipation of the KT66 I use the pentode connected values like:


.9 X 25W / Plate Voltage = Dissipation (W) of anode?


Then to measure Actual RMS output I load the amp with large wattage 8 ohm resistor and scope across it.  Then turn up to the onset of clipping and use V X V / 8 ohm = RMS output?


Which leaves me with the preamp voltage question.   I have preamp, driver, CF, stages but they won't really drive the KT66 with any reasonable volume.  So I need another stage which is why I wanted to go ECC832 ( Hi mu section for the driver and Low mu stage for after the TS to bring it back up plus some added gain.  I also figured that the impedance would be a bit lower going into the KT which i was hoping would help with blocking and such at high levels.  I knew I would need a gain stage with some sort of voltage divider between the TS and KT66 grid.  Just unsure of how to get a decent lively gain out of the preamp and not smash the KT with to much AC.


Thanks everybody.

[shooter]

stages but they won't really drive the KT66 with any reasonable volume.

Cathode resistor Vdrop: 27.8v

I use Vk as my "ballpark" clean drive goal, you want at least that much Vrms as a drive.  If you're not getting that much, play with your CF and change to a gain stage

.9 X 25W / Plate Voltage = Dissipation (W) of anode?

Yup, I shoot for anything from 80 - 105%, it's a jugging game with Rk, and plate volts.

[goldstache]

Can someone direct me to some info on designing the 2nd stage of a preamp.  Say after my first stage I have 5vac out from a 250mv/1k sine input.  How do I go about plotting the second stage? 

First stage is a real nice looking sine wave with no clipping.  Just not sure how to plot the load with that kind of input. What kind of swing can the grid of the following stage handle if it's near center biased like the first?

Thanks for the help.  Trying to get a handle on what's going on stage by stage.     

[SILVERGUN]

Here's some good stuff...
http://www.valvewizard.co.uk/gainstage.html
 
https://vimeo.com/33244955

[goldstache]

Silvergun, most helpful friend!

I was getting some saturation, as the negative troughs were clipped. 

[DummyLoad]

a decent choice for 6L6GC/KT66 SE OT. use 8K tap  6L6 with ~~400V b+.


http://www.classictone.net/40-18031.pdf


--pete

[goldstache]

a decent choice for 6L6GC/KT66 SE OT. use 8K tap  6L6 with ~~400V b+.

Pete can you expound on the calculations for the primary.

I was going with valve wizards Va X Va / Pa to achieve a center bias point

My numbers:
Va 366V (after subtracting Vk)
Pa 25W (KT66 Pentode Mode)

Then,
366V X 366V / 25 = 5,358 or around 5K.

I have a 5K/16R @15W OT on it currently. 
I do have a 8K/8R @15W I could try. 

Just wondering how you arrived at the 8K primary.  Experience or calculation?

Thanks a lot.  I'm gonna get it dialed in with everyone's help.

[DummyLoad]

i was assuming 450V+ (425V -Vk) for b+ and i assumed ~25W Pdiss so ~7.2K - used ohm's law like you did. if you want a 15W SE OT, that one is hard to beat for the price. i should have read the reply where you posted voltages. apologies for the confusion.


--Pete

[shooter]

366V X 366V / 25 = 5,358 or around 5K.

I typically fall about where you're at, what I'm finding, the lower the plate V, the less clean signal,  I hope to get my next breadboard set closer to DL's #s.  Again, it's a juggle Rk Vp, play, juggle Rk Vp, play....

[DummyLoad]

built a 2 x EL34 SE with 25W edcor OT - 3.5k with ~410V b+ with Pdiss around 22W per pentode. this was done on a breadboard. the 2K g2 resistors took some of the edge off when pushed hard, however, IIRC 1K parts worked fine if you backed off from full up.


attached are the plans that i was going to build. it was to be built in a hoffman ver.2 stout chassis.



--pete

[goldstache]

Interesting input!  Thanks, I may have to commit it to a build to finish it up as the breadboard has some oscillations when maxed.  I ended up splitting the CF cathode load in half and tapping an output to the tonestack.  I also ended up tapping off a split anode load for the last stage of the preamp.  I was getting all sorts of oscillations at full up and it was destroying signal and I really couldn't tell what it sounded like full throttle.  At this point in the tweaking, I'm fighting the breadboard.  So it's time to put it in a chassis and walk it to the finish line. 

I checked AC voltage to the KT66 and it's far below VK's voltage.  So once it's a proper amp, all my split loading may go and I can actually hear what the amp sounds like cranked up.  I should know better than to make adjustments to polish turds that are brought on by lead dress on the breadboard. 

Time to drill some holes!
Thanks everyone, I will update once its all together. 
I'm really digging the 4 input topology on a single ended.  It isn't a Marshall, but it's got that kind of tone shaping and gain characteristic.

Dummy Load:  That schemo looks cool!  Really dig the parallel stage to drive the para EL's.  Makes sense from a load perspective and would prolly serve you well to avoid blocking and such. Neat stuff!

[goldstache]

Before dropping it a pan, I wired up my IRF820 CF where a normal 12ax7 CF was. 

On the 12ax7 CF I was set up just like a marshall only changes were Split Cathode load resistors.  This reduced signal into the tonestack and made the controls more usable IMHO. So when I supplanted the IRF820 in that spot, I went with similar topology.  I just added a 100R/1W resistor from the plate of the previous stage to the Gate of the IRF820.  I also added the 12V Zener protection diode, anode to source, cathode to gate. Attached is a drawing with voltages. 

How do I bias this thing in with DMM or do I have drag out the scope?

Thanks.

[sluckey]

I think all you have to do is just turn it on, plug in, and play.

[goldstache]

I did.  And surely it worked.  My Gate and Source at the same voltage.  I've read somewhere that the Gate should be around 1/2 B+.  Which I don't understand as it is fed by the previous plate voltage. 

Anyone have any pointers, is the split cathode load acceptable?

[PRR]

> Attached is a drawing with voltages.

If those are the voltages, it clearly is not built as drawn.

> How do I bias this thing

Same as most small stages. You pick bias resistors which will set voltages to reasonable values, pick resistors to pass reasonable currents. (Yes, it is circular.)

You know the 12AX7 with G1 near zero, 2.2K and 100K will set its Plate in the upper half of the B+ range. The FET Follower will, uh, follow: Source sits a couple volts higher than Gate which is at tube plate voltage. If the tube plate can swing, the FET Source can swing. The FET will pass 223V across 44K or 5mA. Hungry, so watch (don't touch) FET heat. There's also 223V*0.005A or 1.13 Watts in two resistors, so half-Watt parts are only good for 29-day warranty, use 1W, 2W if handy.

> the Gate should be around 1/2 B+.

That's for absolute maximum swing with unlimited input.

Here it can only swing what the tube gives it. Looks like 90V up-swing, though the tube is going to cut-off and doesn't want to swing way high. The FET can easily follow that. For Ab-Max swing, you'd pick a lower cathode resistor (even 470r) to get the Plate voltage down nearer half. Except 12AX7 with 100K load can't down-swing very far. Mid-swing computes to 230V, and 223V=230V for practical purpose.

> is the split cathode load acceptable?

If that is what you want. You throw-away HALF your hard-earned Gain and Swing. But in guitar amplifier, that is often a Good Thing. You gain a well-defined output impedance: 11.1K clean, 22K when FET cuts-off, which limits interactions and bias-shift in tone caps.

[PRR]

> My Gate and Source at the same voltage.

Gate is a couple volts higher than Source.

However gate is plate which is a high-ish impedance, Source is a near zero impedance. Meter loading drops gate-plate voltage more than Source voltage. If you have two meters, you will see Source sag as you poke at Gate. Interesting but not important.

[goldstache]

Thanks for the help.  I will up my source resistors' wattage value. 

Interesting about the swing difference from hollow to solid state.  A bit over my head, lots to think about.

Attached is my updated schematic.  Anything look funny to anybody?

One thought I had was, the Split plate load on V2B helps me shave some voltage off the signal headed to the KT66.  But does it also hurt me in the output/input impedance heading to the power tube? 

It's close enough on the breadboard that this is the schematic I'm gonna build from and tweak it from there.   

thanks again. 

[goldstache]

Well it's all built up and as I suspected the noise was from the breadboard.  It's all as predicted.......... until.

I got some Genelex tubes laying around and wanted to audition them. 

I popped in the KT66 into the amp and noticed that the cathode voltage crept up 5v.  It was close to 25 volts on the  cathode with several other brands of KT66's.  What gives? I thought cathode bias was supposed to auto bias with in a certain degree. 

Having 30VDC on the cathode moves me up to about 26W dissipation, was at around 22W dissipation. 

The funny thing is that my Screen and Plate voltages are only a few volts different. 

If I want to use the Genelex do I have to lower my bias range so it will accommodate this brand as well as the others I tried that drew less current?

The amp rips and sounds way louder than I thought it would.  All the iron stays cool even after being played and idling for a few hours. 

How to proceed?

Thanks

[shooter]

What gives?

you'll have to wait for the theory, but I just ran 7 6K6's through my breadboard and they all varied, both in gain and bias point.  Ya gotta love the Analog world

[DummyLoad]

just had a close look at your schematic: why the tube rectifier? if hum and buzz are an issue, then solid state rectifiers and larger reservoirs are the path one should take with SE. the current doesn't vary much in SE, so there's no timbre in notes with SE from sag like there is with push-pull with tube a rectifier.


replace Rk with a 390/400/430.



respectfully,


--pete

[goldstache]

I agree with the rectifier not being necessary, but my buddy who I'm building it for really wanted it to be like a mini JTM.  I tried to talk sense to him, as I believe you're right.  I will go with a larger Resistor just to be sure. 


Thanks