Second Amp Design

[Swampertech]

Hello fellow tubeheads! Please take a look at the two schematics I have included. The circuitry is based on, of course, Leo’s designs, sluckey and Rob Rob modified circuits, and a few other recommendations from “the experts” on this and other sites. It will be housed in a PR chassis and tweed cabinet with 12” Celestion Gold speaker. I included both Fender and Marshall voiced input preamps (Rob Rob normal/lead channels), cathode follower TMB tone control, a one tube transformer-less reverb circuit (channelroadamps), a one triode tremolo (sluckey Tremor-Lator design), FX in/out, MV, LTP PI w/selectable FB/no FB switching, fixed/cathode bias switching (Rob Rob), KT66 power tubes for high power or switch the rectifier to a 5Y3 and install 6V6 power tubes for lo power, and a 10% power switch for practicing (Rob Rob). I’ve tried to allow for the ability to adjust overdrive in the preamp sections with just a change of summing resistors. I fused the B+ in the event of a shorted tube, and added protection diodes to the rectifier tube.

As this will be only my 2nd build, I’d appreciate any input you may offer. Thanks for taking the time.

[shooter]

the two schematics I have included.

I only c 1, which is fine, since I would start there anyway, make it work using a known GOOD pre stage, then build the pre.
The biggest problem with big designs is gremlin creep.  By building in "stages" you catch n eliminate the creepin G's b4 they can take over 

[sluckey]

Pwr amp looks good. Where's the preamp?

[Swampertech]

Sorry, I didn't get the preamp included. My bad!!

[Swampertech]

the two schematics I have included.

I only c 1, which is fine, since I would start there anyway, make it work using a known GOOD pre stage, then build the pre.
The biggest problem with big designs is gremlin creep.  By building in "stages" you catch n eliminate the creepin G's b4 they can take over 

Hey Shooter, I had initially contacted Randall Aiken concerning some of the circuits and values, and he suggested I learn LT Spice and run a simulation. I wish I had the time and software background to do that but...............
Most of the circuits have been added based on known good amp designs. I guess the main concern has been (1) will the 1 tube transformer less reverb work?) and (2) will the Tremor-Later wobble the signal enough in it's present configuration?) I guess I will find out, unless you know someone I could pay to run a Spice simulation of the whole amp!
Anyhow, I'm excited to start this project and have just about completed the BOM. Most of the components I will use will be 1-5% resistors, Auricaps in the signal stages, F&T filter caps, and Heyboer transformers. I have tried to list the wattage ratings of most resistors, but do have a question concerning the ratings of the tail resistors.
Thanks for your input!!

[shooter]

the main concern

look in the achieves, Tubnit has a great thread on ALL things verb.
From all the FB on Sluckeys Trem version, It will wobble about anything you put it in more better than most/any stock.

If I boiled ALL the great help/knowledge gained here it could be summed up;
GO SLOW, methodical, logical, test EACH stage.  ASK when "it doesn't seem right"
enjoy
 

[Tony Bones]

It's REALLY, really helpful to test new designs a piece at a time, but it's still merely really helpful to test new builds (of a proven design) in pieces. Unless you're an experienced builder, the odds are very low that you'll build your amp and it will work well first try. Not because there's anything wrong with the schematic...

I think what shooter is suggesting is that you build the amp in sections and test (then troubleshoot) each section to whatever extent you can. Not everyone does it that way. Lots of people build the whole thing and plug it in only to find that it doesn't work. Troubleshooting then depends largely on isolating the section that's not working and the difficulty is multiplied by the number issues. It's just easier when the unknown section is smaller to begin with.

[Swampertech]

It's REALLY, really helpful to test new designs a piece at a time, but it's still merely really helpful to test new builds (of a proven design) in pieces. Unless you're an experienced builder, the odds are very low that you'll build your amp and it will work well first try. Not because there's anything wrong with the schematic...

I think what shooter is suggesting is that you build the amp in sections and test (then troubleshoot) each section to whatever extent you can. Not everyone does it that way. Lots of people build the whole thing and plug it in only to find that it doesn't work. Troubleshooting then depends largely on isolating the section that's not working and the difficulty is multiplied by the number issues. It's just easier when the unknown section is smaller to begin with.

Thanks TB, I do have this old BF Twin PT that I'm planning to use to build a bench power supply just for the purpose you suggest. And you're right, it may work in another amp, but not guaranteed to work in this configuration.

[shooter]

just for the purpose you suggest
 
the 1st "piece" I test, and usually on the BB IS the PS.
I calc the ~ load for each tap, hunt up/order an R that gets me close to that taps load.
by doing that I can tell better the sag, where B+ will really land etc.
once that's done, install in amp
PI /PA next (well I do xSE so driver/volume + PA)
pre last
everybody together now, SWEET 

[Swampertech]

just for the purpose you suggest
 
the 1st "piece" I test, and usually on the BB IS the PS.
I calc the ~ load for each tap, hunt up/order an R that gets me close to that taps load.
by doing that I can tell better the sag, where B+ will really land etc.
once that's done, install in amp
PI /PA next (well I do xSE so driver/volume + PA)
pre last
everybody together now, SWEET 

Great advice! SWEET!!

[2deaf]

Reverb

The 6GH8A pentode driver looks like it will overdrive with an input in the 1.5V_p to 2V_p range.  You are walloping it with way more than that.  R15 (!M) needs to be changed to a voltage divider.

R16 (68K) does nothing until the frequencies get out there towards light because pentodes have very little Miller effect.  You can put a capacitor from pin 2 to ground to roll-off highs, maybe as low as 3KHz for reverb purposes.  To only prevent high frequency oscillation, R16 would probably need to be greater than 220K.

R18 (15K) needs to be 3W minimum.

The entire 6GH8A takes down about 14mA (what a pig).  This will cause a voltage drop across R58 (4.7K) of around 100V.  You don't want that, so connect the pentode to B2 where it won't much affect the supply node voltages.  You are going to want the most filtered supply for the triode and it is taking down less than 2mA, so connect it to B4.     

[2deaf]

Connect V4B (tremolo) to the B2 power supply node.  Now R55 (220K), R56 (220K), R57 (1K) and R58 (4.7K) can all be 1W.  The choke seems a little overkill to me.

I don't like the bias circuit.  What bias voltage do you figure for KT66's under these conditions?

[2deaf]

Effect Loop

Sleeves need to be grounded on the Send and Return jacks.

If the Send/Return is going to function as an effects loop, VR1 and/or VR2 are going to have to be turned way down in order to get an effects level signal at the Send jack.  You may not like that.  Replace VR9 (MV 1M) with a 220K resistor and move the MV (VR9) so that it is being feed by R28 (220K) and the wiper is going to the tip of the Send jack.  You can eliminate R28 (220K) if you want.  Now the send level is controlled by the MV allowing you to adjust VR1 and VR2 to taste.

Most effect devices take a relatively small signal at their inputs and output a signal that is similar in magnitude.  So as far as the PI is concerned, it is getting a signal that is the same as having the MV turned way down.  You may not like that.

[2deaf]

Your convention on the potentiometers seems to be that the lower end on the schematics corresponds to the left lug when looking at the back of an actual potentiometer.  If this is the case, the Mid pot is wired backwards -- midrange frequencies will decrease as the numbers get larger on the control knob. 

I would eliminate R8 (6.8K) so that the midrange can be turned clear off.

[shooter]

so connect it to B4.   

my brains already spinnin mud trying to catch up to ya 

Are there a few to many low signal tubes plugged into common tap(4) 
thinking maybe V1 needs a tap 5?

[Swampertech]

Reverb

The 6GH8A pentode driver looks like it will overdrive with an input in the 1.5V_p to 2V_p range.  You are walloping it with way more than that.  R15 (!M) needs to be changed to a voltage divider.

I have seen that 1M resistor replaced with a 1M pot, the wiper feeding the grid. Would that be enough to tame the signal coming from the output of the tone control?

R16 (68K) does nothing until the frequencies get out there towards light because pentodes have very little Miller effect.  You can put a capacitor from pin 2 to ground to roll-off highs, maybe as low as 3KHz for reverb purposes.  To only prevent high frequency oscillation, R16 would probably need to be greater than 220K.

So the chances of oscillation are slim, but there still needs to be a large (>220K) grid stopper?

R18 (15K) needs to be 3W minimum.

I believe you use a 5W in your 12DW7 design, right?

The entire 6GH8A takes down about 14mA (what a pig).  This will cause a voltage drop across R58 (4.7K) of around 100V.  You don't want that, so connect the pentode to B2 where it won't much affect the supply node voltages.  You are going to want the most filtered supply for the triode and it is taking down less than 2mA, so connect it to B4.   

So 2deaf, would not your 12DW7 transformerless reverb design work in this application with this tank? The 12DW7 has one section that can drive with 10.5mA, which should be enough to drive the reverb tank input. I read that channelroadsamplifiers state that "According to the Accutronics data, this tank (4EB2C1B) can be fully driven with a signal of 3.2 mA RMS ≈ 4.5 mA peak.

[2deaf]

I wasn't trying to talk you out of the 6GH8A driver, just giving you a heads-up about the current consumption.  The 12DW7 driver takes a good-sized bite of current, also.

Accutronics idea of fully driven and everybody else's idea are significantly different.  If a driver won't do at least twice Accutronics value, then I won't consider it.  The 6GH8A driver should easily do more than twice the Accutronics value.

[Swampertech]

I wasn't trying to talk you out of the 6GH8A driver, just giving you a heads-up about the current consumption.  The 12DW7 driver takes a good-sized bite of current, also.

2deaf, the only reason I wanted to use the channelroadsamplifier 6GH8 circuit was that I was trying to incorporate a one tube reverb (using a PR chassis - have to add one tube socket hole). I saw Tubenit's one tube reverb circuits on this site, but most still require the use of a transformer. That's something I'd like to eliminate. Your 12DW7 transformerless reverb provides one triode that delivers 10.5 mA, over twice what Accutronics requires, and the triode for the recovery circuit. The 6GH8A tube draws about 25-30mA (300 peak), which I took into account when I researched the power transformer.

Accutronics idea of fully driven and everybody else's idea are significantly different.  If a driver won't do at least twice Accutronics value, then I won't consider it.  The 6GH8A driver should easily do more than twice the Accutronics value.

[2deaf]

The 6GH8A tube draws about 25-30mA (300 peak), which I took into account when I researched the power transformer.

I figure the 6GH8A pentode will have a plate current around 9mA with a 300V supply.  The screen grid looks like it takes about a third of the plate current value, so 3mA.  The total current consumption for the pentode section would then be 12mA.  The triode looks like it will idle between 1.5mA and 2.0mA, so I rounded it up to 2mA to get a total 6GH8A consumption of 14mA.

The datasheet shows a maximum DC cathode current of 20mA for the pentode, but that's not what it idles at here.  The idle current is what we are concerned with for power supply purposes.  The 300V peak is an AC value and is of no concern here.

[Swampertech]

So, 14mA is not going to tax my power supply is it?

Selected transformer voltage: 345-0-345
Calculated voltage at first capacitor (B+): 469.2V. Subtract 6V, if using choke.
Calculated filament current (typically the 6.3v secondary): 3.92A
Calculated current: 125.25mA at 8806R calculated load (10% plus factored in). Preamp valves current draw is estimated at typical 12AX7 max dissipation of 1.2W at 330V, i.e. 3.6mA per triode

[2deaf]

So, 14mA is not going to tax my power supply is it?

No.

[Swampertech]

Effect Loop

Sleeves need to be grounded on the Send and Return jacks.

If the Send/Return is going to function as an effects loop, VR1 and/or VR2 are going to have to be turned way down in order to get an effects level signal at the Send jack.  You may not like that.  Replace VR9 (MV 1M) with a 220K resistor and move the MV (VR9) so that it is being feed by R28 (220K) and the wiper is going to the tip of the Send jack.  You can eliminate R28 (220K) if you want.  Now the send level is controlled by the MV allowing you to adjust VR1 and VR2 to taste.

Most effect devices take a relatively small signal at their inputs and output a signal that is similar in magnitude.  So as far as the PI is concerned, it is getting a signal that is the same as having the MV turned way down.  You may not like that.

2deaf, I have been considering the Metropoulos FX loop, which should solve any loading or gain issues.

[Swampertech]

Connect V4B (tremolo) to the B2 power supply node.  Now R55 (220K), R56 (220K), R57 (1K) and R58 (4.7K) can all be 1W.  The choke seems a little overkill to me.

I'm nixing the choke, so that there is no flyback issues.

I don't like the bias circuit.  What bias voltage do you figure for KT66's under these conditions?

I read that the KT66 in ultra linear mode would need -50 to -65 VDC for bias. Could I just pick the bias voltage from the 340-0-340 HT secondary tap and rectify/filter that for bias?

[sluckey]

Could I just pick the bias voltage from the 340-0-340 HT secondary tap and rectify/filter that for bias?

yes

[Swampertech]

2deaf, sluckey, shooter - Thanks guys for your valued input. I thought that since I'd used 2deaf's mods to sluckey's Tremor-Lator in this design, I would go ahead and insert 2deafs 1 tube transformer less reverb. I would run the driver side of the 12DW7 at +B2 level, and the recovery triode of the 12DW7 at +B4. Please see attached schematic.

[shooter]

at +B2 level

schematic still reflects +B4 for both
I'm still Leary of the +B4 tab with (4?) tubes.  NO proof, just don't want you having to track down hum or coupled positive FB into V1

[Swampertech]

Shooter, yes I know. I just didn't change that when I substituted 2deaf's reverb. When I can get finalized on the circuits, I'm going to nail down the PS, and may add a +B5 as you suggested. Again, thanks for your suggestions.

[2deaf]

I read that the KT66 in ultra linear mode would need -50 to -65 VDC for bias. Could I just pick the bias voltage from the 340-0-340 HT secondary tap and rectify/filter that for bias?

Like sluckey said, yes.  The 50V bias tap should work, also.  Attached is a suggestion.

[2deaf]

2deaf, I have been considering the Metropoulos FX loop, which should solve any loading or gain issues.

The Metropoulos loop is well designed and built and was designed primarily for that position.

[Swampertech]

The choke seems a little overkill to me.

2deaf, in most of the amps I've seen there is a choke between the +B1 and +B2 nodes, which seems to drop the screen voltage down to a usable level- a bit lower than +B1. Can you explain your "overkill" reasoning? Thanks!!

[shooter]

I'll answer for me 
I have 2 11lb'rs I use for book-ends 
(possibly use in an audio build, maybe)

the non technical reason for me;
letting G2 wander around within limits adds tonal results - player preference kinda thing.
I like a stiff, no tone PS so I build with UL OT's and lots of big Uf Ecaps in the PS.
Weight!  I like head units with a smallish footprint, 11lbs is a lot of real-estate and weight.

[Swampertech]

I'll answer for me 
I have 2 11lb'rs I use for book-ends 
(possibly use in an audio build, maybe)

the non technical reason for me;
letting G2 wander around within limits adds tonal results - player preference kinda thing.
I like a stiff, no tone PS so I build with UL OT's and lots of big Uf Ecaps in the PS.
Weight!  I like head units with a smallish footprint, 11lbs is a lot of real-estate and weight.

Shooter, so you just allow the screen resistors to regulate the screen voltage? Typically, I've seen a 1K between +B1 and +B2, and the output of the 1K would feed the 470 ohm screen resistors. The KT66 spec sheet says typical push pull voltages would be 450V-plate and 400V-screen.

[2deaf]

2deaf, in most of the amps I've seen there is a choke between the +B1 and +B2 nodes, which seems to drop the screen voltage down to a usable level- a bit lower than +B1. Can you explain your "overkill" reasoning? Thanks!!

194B at 4H and 90mA is plenty for two big bottles.  Save some space, weight, and money.

[Swampertech]

sluckey, 2deaf, and all.

Hopefully I can post these 2 new schematics. On the first one, pre amp section, I’ve substituted 2deaf’s one tube transformer less reverb, reassigned the +B voltages - adding a +B5 for the following tubes:

+B5 - V1A, V1B, V2A, V2B, V3B, V4A - should draw 7.2 mA using 1.2mA/triode

The rest of the tubes I’ve assigned to these PS nodes:

+B4 - V3A, V4B - 13.2mA

+B3 - V5A, V5B - 2.4mA

+B2 - V6 (Screen), V7(Screen) - 18mA

+B1 - V6 (Plate), V7 (Plate) - 124mA

The current draws listed are max according to the tube manuals, but I’d invite your real world input on these calculations. I also corrected the mid pot wiper connection, based on 2deaf’s recommendation, along with nixing the 6.8K R8.

The second schematic is also a work in progress, as I’m installing a Metropoulos FX send/return circuit but haven’t removed the FX in/out jacks. I changed the cathode resistors of the KT66’s to agree with the tube specification sheet. Also bumped the cathode bypass caps up to 50uF. Another change, based on 2deaf’s recommendation, is the bias circuit. Another change is the elimination of the filter choke. And, as you can see, the +B smoothing filters are not correct.

But other than that, it’s getting close to bench testing each circuit. I know you guys can just look at this collection of circuits and see what’s wrong. Thanks for your input as always. I know this is an ambitious project, but please forgive a newly for his enthusiasm.

[shooter]

allow the screen resistors

In my personal builds no, I use a UL OT, which is kinda sorta a G2 choke.
For paying musicians they "tend" to like just G2 R "sound"
G2 R limits current keeping it ~ safe.

It's one of those nuanced things that really needs to be played verses analyzed.
And by the time n effort you're putting in, within a year you'll have "Your magic sauce" pretty well dialed 
 

[Swampertech]

allow the screen resistors

In my personal builds no, I use a UL OT, which is kinda sorta a G2 choke.
For paying musicians they "tend" to like just G2 R "sound"
G2 R limits current keeping it ~ safe.

Dr. Z has a KT66 amp using UL OT. I'll look some more at UL.

It's one of those nuanced things that really needs to be played verses analyzed.
And by the time n effort you're putting in, within a year you'll have "Your magic sauce" pretty well dialed 

Oh, I'm open to any suggestions which improve the stability. A UL OT would eliminate one of the +B nodes, and the KT 66 has been used with UL OT for a while. My choice of the Heyboer TO-26 was based on my first build (5E3).

[Swampertech]

Shooter, here are some more realistic PS dropping resistors. +B1-450V, +B2-410V, +B3-335V, +B4-315V, +B5-308V. Sound about right for the circuit?

[shooter]

It's not about the voltage per tap per se', it's about having multiple tubes with varying signal amplitudes.  I like V1 by itself, that way tubes downsteam won't be tempted to induce FB back into V1.

leave your grid R's, build the amp, things like UL or clamped G2 voltages are just "flavor".  Once you get bored with an amp, experiment