New push pull triode amp - Tubeless inverter

[RJ Guitars]

Hey guys,

Edit 8/19/07 - Updated schematic and new photos in a later post

Edit 8/17/07 - Sound Clip added... see disclaimer in a later post
http://home.comcast.net/~rjguitars/Musi ... e_Rock.mp3

Edit 8/15/07 - Schematic Added - Clips soon!

I'm just finishing up an all triode amp using an interstage transformer for the phase inverter... no inverter tube needed. The interstage transformer was $12 direct from Edcor.

Since this inverter idea was new to me I studied all I could find on inverters and it turns out you can't make a simpler phase inverter than this. I found several good resources and even some other guitar amps that use this idea... http://www.audioxpress.com/resource/aud ... a200ac.pdf

Both Fender and Gibson made amps with transformer phase inverters... I like the simplicity of this Fender schematic... http://www.ampwares.com/ffg/schem/music ... _schem.jpg

My circuit is a three tube amp with a 12AX7 preamp, a post inverter 12AU7 driver, and an ECC99 output - All triodes.

I'm cleaning up my drawings and I haven't had a chance to make a sound clip yet... I'll share with you as these become available.

It is surprisingly loud for only a few watts - yet it is not painfully loud.

This new amp sounds really cool. I can even get it to feedback with my Les Paul - not a simple task with less than 5 watts.

Not a lot of clean headroom - but great sound and an especially fun crunch tone. I can also get a very "Dumble like" singing sustain when I've got it cranked...

This was fun!

rj

[jelle]

That's a cool approach!

SOUNDCLIPS>>>>SOUNDCLIPS!!!

Jelle

[skyboltone]

I like it. I'm kind of a oddball amp guy myself. What's the front panel made from?

[VacuumVoodoo]

My circuit is a three tube amp with a 12AX7 preamp, a post inverter 12AU7 driver, and an ECC99 output - All triodes.

I'm not sure I understand correctly that you've put the 12AT7 after the transformer phase inverter so it drives the ECC99?
I would rather use both triodes in 12AT7 in parallel to drive the PI transformer primary. The transformer PI will have enough juice in it to drive the ECC99 (or most any other power tube) push-pull into AB2. This could give a pretty beefy output stage overload/compression.

BTW. "the Audio Classroom" series by Crowhurst you linked to is excellent , should be required reading.

[RJ Guitars]

I like it. I'm kind of a oddball amp guy myself. What's the front panel made from?

Long Story... but it's basically a reverse image on a poly acetate sheet (overhead viewgraph) then I sprayed gold paint behind it. For reasons unknown to me... some chemical reaction occurred and the gold turned antique... it came out kinda cool & retro but not by design.

rj

[RJ Guitars]

My circuit is a three tube amp with a 12AX7 preamp, a post inverter 12AU7 driver, and an ECC99 output - All triodes.


I'm not sure I understand correctly that you've put the 12AT7 after the transformer phase inverter so it drives the ECC99?
I would rather use both triodes in 12AT7 in parallel to drive the PI transformer primary. The transformer PI will have enough juice in it to drive the ECC99 (or most any other power tube) push-pull into AB2. This could give a pretty beefy output stage overload/compression.

Oh yeah that's what we did (I'll try to get a schematic out tonight). I'm not sure I'd do it exactly the same again myself... but it has some really interesting characteristics. As an experimental amp it is wonderfully successful. I've seen a couple amps with post inverter triode drivers and the idea has some merit. I believe it helps shape the tone a bit.

I was trying to get some sound clips last night and I found that although I get really great sustain (volume most of the way up) and there are a couple nice sweet spots... at full crank it's got this really spongy compression that I got bored with pretty fast... Europa kinda leads are fun but I have to back the gain down to do anything else.

As a simple amp build, the tubeless inverter allows you a push pull amp using just two tubes, I might recommend that build for a first timer.

Amen to the Crowhurst article - it's great stuff!

rj

[RJ Guitars]

Hello Amp Builders - I appreciate your comment on this project.

I have added a schematic and if you'll notice I give credit to the originator of the design - Jason Cox of Boozhound Labs (not me). He has been a mentor in training me in tube amp building. He's a modest guy that shares his vast knowledge freely. He always has a novel idea that he'll add to your project... maybe he will jump in and comment on the how and why of this design?

For anyone interested - I've also got a pretty decent layout for this amp and I'm working on a parts list.

Again thanks for the comments... I've recorded a few sound clips and I'm working on editing them now.

(Schematic is in the original post)

rj

[lastwinj]

its called an interstage transformer.

germ

[PRR]

V1+V2 could implement a full guitar amp without the interstage iron. Drive V2a grid from V1b, and ground V2b grid. The signals will be obviously unbalanced at V2 plates, but the cathode resistor on V3 will tend to give pretty good balance on V3 plates. Or jack-up V2 grids ala Fender longtail.

Transformers have low impedance compared to grids or even 12AX7 plates. I would not drive the tranny from a 1Meg pot: at half-volume the pot presents about 250K to the transformer, which on paper will clobber its frequency response.

The 0.01uFd coupling cap here is low for a transformer. Will tend to give a midbass peak and bass droop.

These "flaws" may not be bad for a small guitar amp.

The on-paper "best" design might be to parallel V2, drive it from the pot, load it in ~27K, use 0.5uFd to drive the transformer, and the transformer drives the V3 power tube grids. A single unit of 12AU7 is a classic interstage transformer driver. Two units of 12AT7 give near the same source impedance with higher Mu.

Fender Musicmaster Bass is the same idea done bigger and cheaper.

Are you really using the Edcor $12 interstage? That's a heavy load for any small tube. OTOH the 2.5W rating implies it won't saturate with any signal level you can easily put into/through it. So it should be happy at the power tube grids, even for triodes needing more drive than ECC99.

You can go even futhur, eliminate most of the Rs and Cs in an audio amp. Nobody (except a few hi-fi-natics) has done this in decades..... blast from past.

[RJ Guitars]

Transformers have low impedance compared to grids or even 12AX7 plates. I would not drive the tranny from a 1Meg pot: at half-volume the pot presents about 250K to the transformer, which on paper will clobber its frequency response.

The 0.01uFd coupling cap here is low for a transformer. Will tend to give a midbass peak and bass droop.

PRR you are on to something here... I've talked about this being a fun project and it does indeed produce some cool sounds... I haven't brought up any of it's failings yet. Keeping in mind that the driving idea behind this build is to create something that sounds big yet is not especially loud, I'm over 90% of the way there... yet, I don't think I've built the ultimate amp yet.

Failing #1 - I can find a sweet spot or two but I'm getting a very bright tone out of the OD input and a very dull tone out of the Normal input. Using the OD channel, if I mix just the right amount between the two volume controls, I can find those sweet spots. I don't think the Normal channel is very usable as it is since I have to use the bridge pickup on full treble to get any reasonable tone and even then it's not real bright.

As it stands, I'm kinda using the normal pot as a master volume and to adjust in some bass to the overly bright OD channel. It works O.K. although I don't think this is a good permanent design. (If I remove the bypass cap from around the OD volume pot I can kill that brightness, so I'm not worried about that for now.)

I'm thinking I've made some sorta high pass filter to ground by going out of the Normal volume pot into the interstage transformer and I think I need something different here. I was thinking of moving the master volume feature onto the output tubes... any comments?

BTW, did anyone notice that this amp only goes up to 6, it's not very loud.

(we'll save the other failings for discussion later).

rj

[PRR]

> very bright tone out of the OD input and a very dull tone out of the Normal input.

Could be.

There's "conventional" and "different". This input scheme is "different". I wouldn't think to do anything like that. I've assumed it was based on someone's experience and testing, and overlooked the odd parts. Being unconventional, it isn't easy to analyze at a glance. (One reason "conventional" is common is that good engineers are lazy and like simple things.)

At a glance, Normal will be somewhat mellow, and OD will be gutless for some pot settings, but I don't see huge difference. Probably I'm overlooking something.

> create something that sounds big yet is not especially loud,

You got a lot of "small parts". 0.01uFd coupling caps. 1Meg (small conductance) pots. 5ufd cathode cap. If "big" involves "bass", this isn't a big sound.

Is this the Edcor "15K" transformer? Assuming they mean 15K at 20Hz, then when the second pot is at halfway the 15K tranny sees 250K source and infinite load. Inductive reactance is going to shift the bass cutoff to 333Hz. Full-up we see the V1b plate, about 40K, F3 near 60Hz. The 0.01uFd cap resonates with the tranny near 500Hz. Depending on pot setting you could have a mild midbass resonance or a long slope-off.

Both pots are tone controls. I'd assumed the big interactions were deliberate, based on trial. It's a real pain to analyze; one reason "conventional" designs are conventional is they tend to less interaction and simpler analysis.

> something that sounds big

Aside from fuller-range circuit constants: there are ear effects. The ear's bass response and nonlinearity are very different at 100dB SPL than at 85dB SPL. You just won't get the chest-smack of a BIG amp with anything less. Your ear can tell the size of a source, and won't confuse a 4-inch for a Marshall Stack. My thinking is that low-power amplifiers with big dreams should use the biggest possible cone(s). Compare a Champ 8" against the same Champ into a Marshall Stack. Not only will it go louder, it stays big at lower power levels. But for best effect you may need to upgrade the Champ's bass response: it was balanced bass-shy so an Eight Inch would not slap bad, with just a 12" you can and should use more iron.

> I was thinking of moving the master volume feature onto the output tubes... any comments?

When I look at that and move anything, I want to move EVERYthing. If you think this plan has promise, hot-up the iron and start tweaking. If you like the transformer splitter, I'd start with a known-good plan like that BassMaster or the oldest Gibsons (the oldest were also P-P Triode output as you are trying). The stock BassMaster is not a breakup monster but you have one more twin-bottle and can easily soup-up.

[RJ Guitars]

Hey Guys, I wasn't sure whether to be proud or embarrassed... but now you know why I haven't quit my day job... and yes I listened to a lot of Lynard Skynard in the days gone by.

http://home.comcast.net/~rjguitars/Musi ... e_Rock.mp3

Disclaimer:
I am not Geetarpicker
I am not Dogears
I do not have a studio
I am not a sound studio engineer ...but I stayed in Holiday Inn Express once
I do not know how to record stuff
Etc...


Thank you for your indulgence... The main idea is to give you a listen to a triode output amp. This recording of course favors the settings where the amp sounds best.

For the drums I used BEATCRAFT software.

For recording and editing I use AUDACITY software.

I recorded the rhythm tracks using a direct line out (see my schematic) straight into my computer. It seems to work well but mostly for the clean and semi clean sounds. There are two separate tracks with two similar settings. One is on the right side, the other is on the left.

I recorded the lead tracks (two of them left/right) with a Shure Beta57A microphone into a small "Audio Buddy" preamp and then into the computer.

I used a 10" Fender speaker (ceramic magnet) mounted in an otherwise empty Pro Junior cabinet for all the recordings.

I didn't add any effects and only edited out the significant defects in my playing...

Enjoy,

rj

[toneman]

Quote:RJ Guitars Disclaimer:
I am not Geetarpicker
I am not Dogears
I do not have a studio
I am not a sound studio engineer ...but I stayed in Holiday Inn Express once
I do not know how to record stuff
Etc...... You are not alone, [/img]

[RJ Guitars]

#1 - You got a lot of "small parts". 0.01uFd coupling caps. 1Meg (small conductance) pots. 5ufd cathode cap. If "big" involves "bass", this isn't a big sound.

#2 - Is this the Edcor "15K" transformer? Assuming they mean 15K at 20Hz, then when the second pot is at halfway the 15K tranny sees 250K source and infinite load. Inductive reactance is going to shift the bass cutoff to 333Hz. Full-up we see the V1b plate, about 40K, F3 near 60Hz. The 0.01uFd cap resonates with the tranny near 500Hz. Depending on pot setting you could have a mild midbass resonance or a long slope-off.

#3 - Both pots are tone controls. I'd assumed the big interactions were deliberate, based on trial. It's a real pain to analyze; one reason "conventional" designs are conventional is they tend to less interaction and simpler analysis.

#4 - Aside from fuller-range circuit constants: there are ear effects. The ear's bass response and nonlinearity are very different at 100dB SPL than at 85dB SPL. You just won't get the chest-smack of a BIG amp with anything less. Your ear can tell the size of a source, and won't confuse a 4-inch for a Marshall Stack. My thinking is that low-power amplifiers with big dreams should use the biggest possible cone(s). Compare a Champ 8" against the same Champ into a Marshall Stack. Not only will it go louder, it stays big at lower power levels. But for best effect you may need to upgrade the Champ's bass response: it was balanced bass-shy so an Eight Inch would not slap bad, with just a 12" you can and should use more iron.

#5 - When I look at that and move anything, I want to move EVERYthing. If you think this plan has promise, hot-up the iron and start tweaking. If you like the transformer splitter, I'd start with a known-good plan like that BassMaster or the oldest Gibsons (the oldest were also P-P Triode output as you are trying). The stock BassMaster is not a breakup monster but you have one more twin-bottle and can easily soup-up.

PRR - Again I am greatly pleased that you would share your knowledge and perspective... I've already learned a ton from your your input.

#1 - I want to do something to get that 1meg pot out from in front of the inverter transformer... so when that is settled I'll use your guidance toward getting some "big sound".

#2 - I used the 10K/10K interstage transformed from Edcor - http://www.edcorusa.com/products/transf ... k-10k.html... do you think it's a good choice? Again I'm stunned but pleased at their price on this thing. Makes a transformer potentially cheaper than a tube.

#3 - Yes I agree both pots are tone controls and it isn't entirely random by design, from my perspective it isn't working quite right yet...

#4 - This is a great discussion and maybe something worthy of a thread all to itself. I've had this discussion with a few local guru's lately. One answer could be to add a "Loudness" control like the old home stereo used. Here is some fuel for more discussion - http://www.audioxpress.com/magsdirx/ax/ ... in2734.pdf
Could this explain why the power brake strategy never seems to work well once the amps get into bedroom volume territory?

#5 - I think this unit has plenty of promise and I'm not opposed to learning something even if it doesn't end up being the holy grail of lightweight power & heavyweight tone... It's funny because I didn't find that Fender schematic until I was already trouble shooting this amp... might have saved myself some grief??

Again, I am pleased to learn form you... It's time to go suck some solder fumes and see if I can normalize that normal channel.

rj

[PRR]

> using a direct line out

Much of the charm of a triode output is the way it interacts with the speaker. Some but not all of this is reflected at the speaker terminals (where your line out is). The more accurate recording would be a mike at the speaker. I did similar work for a living and I understand why you skipped that; whipping out a mike enormously complicates a session.

> Makes a transformer potentially cheaper than a tube.

There's cheaper tubes. But they need power and etc. And $4/$12/$20 is fairly modest compared to the value of your labor and the potential enjoyment from DIY. I grant your point.

If what you want is a bedroom amp with transformer coupling....

A classic reason to transformer couple triode power stages is to drive grid current. The only ways to get real power from triodes is high plate and drive voltages, or high grid currents. (Or BIG cathodes and heaters.) Any grid current, even 0.1mA, will drive a capacitor-coupled grid to cut-off, grid blocking. But a transformer handles that fine.

A particular simple form is "zero bias". Use a hi-Mu tube which gives small plate current at zero grid bias. Adjust plate voltage so plate dissipation is tolerable. In push-pull, either way the wave goes, one grid or the other is positive and pulling current. The load on the driver tends to be more constant (less distortion) than a stage which does-or-does-not draw grid current at various grid voltages.

Most LARGE triodes have positive-grid data. Most small ones don't.

12AU7 has positive-grid data, but the grid voltages and total grid drive power tends to be large.

6J6 has positive-grid data and with Mu=38, the grid voltage needed is quite small. The plate current available without grid current is small, but with a step-down transformer you can deliver the small grid voltage and high grid current from a modest-current driver.

And 6J6 is just $4 each!!! It was produced in vast quantity, rarely dies in its normal uses, most of its gear has been obsoleted. It is a bit awkward for audio due to common cathode, but is a really good tube. You could probably score a bakers-dozen for the cost of one ECC99 and a spare.

The condition I see is 100V supply voltage, zero bias, 10Kpp load, around 2 Watts output. Grid drive could come through a 10K:600CT transformer from another $4 6J6, both sides parallel. This bottle needs about 2V drive. Since 6J6 are cheap and you like paralleling, two more in resistor-coupled mode (or one 12AT7) could give enuff gain to come up from guitar pickup and through a lossy tone stack. (You could throw in another interstage transformer, get a little more gain, but not enuff to eliminate a bottle, and frequency response of three cascaded transformers could be less than you want.)

The downside is that 6J6 heater is hungry. Four of them will suck 1.8A at 6.3V. A Fender Champ will do more than double the speaker watts at less than half the heater power. OTOH, 150V-200V power caps today are far cheaper than the 450V caps needed for a Champ.