High Gain Princeton clone - any thoughts?

[Lectroid]

Hi all,

Recently got a small Magnavox chassis with 5y3, 2-6V6s, and a 12AX7.  Inside it was pretty similar to a 6G2 Princeton circuit.  It also had a spare miniature socket.  So I thought I'd add another 12AX7 to the preamp and try to boost the gain of the original Princeton circuit, as well as adding an actual Gain pot to get more control over headroom and distortion.

I have built a couple of small working 6V6 amps but am a near-total noob at tube amp design.  This is is my first attempt to color outside the lines, so...

Anyone have any thoughts about the circuit?  Will it work the way I want it to?  Is this all a huge mistake?  Where did I go wrong?  How to make it better?

All criticisms gratefully accepted.

[PRR]

You probably do not want 1MEG at V2A cathode.

Three high-gain stages in a row on one B+ node is asking for trouble. And you have the parts to separate out another node.

No NFB?

0.1uFd and 220k at power tube grids is 8Hz. Ample for hi-fi. But if you OVER-drive it to grid-blocking it can take a long time to start working again. Don't most similar guitar amps have much smaller grid caps here? (Saves a dime too.)

[Lectroid]

Thanks, PRR.

The 1M resistor was a cut and paste error. Should have read:  1K since I'm trying to overdrive the stage a little here.

I will add another B+ node.  I was just copying off the Princeton 6G2 schematic.  But why exactly is it 'asking for trouble?'  Will it lead to more voltage sag?  Or what?

NFB, also an oversight.  It will be added.

Finally, the 0.1uf/220K combination:

0.1uFd and 220k at power tube grids is 8Hz. Ample for hi-fi.

Those values are a straight steal from the Princeton 6G2 schematic, also found in the 6G3 and 5E3 Deluxes.   I'm certainly open to changing it but I don't quite understand your comment.

What makes it 'ample for hi-fi' is, I assume, the full bandpass across all audio frequencies of interest.  Please correct that if I'm wrong.

But if you OVER-drive it to grid-blocking it can take a long time to start working again. Don't most similar guitar amps have much smaller grid caps here? (Saves a dime too.)

But why might it not work as well in a guitar amp? What's the benefit with raising the frequency of the 3dB cutoff, (which is what I think you're saying.).   For instance going with a .01/220K combination gives a -3dB at 72Hz.  That point strips out very low frequencies I can see, but is there some other point to doing it that I'm not seeing?  What's the main benefit?

I know I'm inferring a lot here, not trying to put words in your mouth.  Just looking for clarity.

By 'grid-blocking', do you mean what Merlin calls 'blocking distortion?'  In rereading that section of his book, he seems to imply that higher value grid stoppers on the 6V6s would prevent it.  Is that a good solution?  Or is there a better one?

Thanks again for your response.  Any further clarifications greatly appreciated.

[sluckey]

Finally, the 0.1uf/220K combination:

Fender used those values in almost every P/P amp back in the late '50s, 60s, and early '70s. A popular mod is to replace the .1µF caps with .047µF caps to reduce the boom boom.

[Willabe]

I will add another B+ node.  I was just copying off the Princeton 6G2 schematic.  But why exactly is it 'asking for trouble?'  Will it lead to more voltage sag?  Or what?

No, not voltage sag. It's 'sneakage' as PRR calls it. The tube stages can/will interact with each other through that single B+ filter cap feeding them. The amp won't be stable.

By 'grid-blocking', do you mean what Merlin calls 'blocking distortion?'  In rereading that section of his book, he seems to imply that higher value grid stoppers on the 6V6s would prevent it.  Is that a good solution?  Or is there a better one?

Yes, same thing.

You can do like PRR said or you can go with large(r) grid stoppers or some combination of both.

Those values are a straight steal from the Princeton 6G2 schematic, also found in the 6G3 and 5E3 Deluxes.   I'm certainly open to changing it but I don't quite understand your comment.

Yes, but, those amps were built to be played with Leo's single coil PUP's, which are not bass heavy. And those amps were not built to be played past 4 or 5 on the amps volume control. I played gigs through a BF Super Reverb with a Strat, no pedals, for years and never turned the bass control past 3, too much bass and that's with single coils. I sometimes got the volume up to 7, mostly 6 or lower, treble bright switch off, middle 7 or 8, treble 7. I played lead on the neck PUP and rhythm on middle. Mine did come with the factory Jensen alnico 10's, most had ceramics and their brighter. Never used the bridge PUP. For me, the mids are too scoped on BF Fenders to use the bridge PUP, on a cranked Marshall, sure.

Find a bunch of hi-gain amp schematics on line and look at how they all keep stripping out the bass between gain stages. Smaller coupling caps, after the CC's signal series R's with a small cap in parallel to let the hi's through and reduce the bottom end, and often smaller K bypass caps so the gain stage is not amplifying the very bottom end as much.   
 

What makes it 'ample for hi-fi' is, I assume, the full bandpass across all audio frequencies of interest.

Yes and do we need to go down to 8Hz on a guitar amp? Even a bass amp doesn't need to go down to 8Hz. It's all mud down there. Most recordings, other than symphonic, the engineers are rolling all the very low end off. That's 1 of the reasons a lot of mics have a low end roll off switch built into the mic for that reason. It's not just for keeping stage rumble coming up through the mic stand out of the PA system.

[PRR]

...I was just copying off the Princeton 6G2 schematic.....

Copy carefully.

6G2, I see 30u-30u-30 with 1k and 10k, for two voltage gain stages.

Your plan, I see 22u-22u-22u-10u with 1k 10k 10k, for three voltage gain stages.

Less filtering up front, perhaps more overall filtering, but all three gain stages on one node. The rule for simple hackers is two sequential stages on one B+ node. Three *may* be stable, but it is a pain to figure; and in this case another node is implied by your added cap.

[Willabe]

And, if you have room, put in a 10K or better yet, 25K or even 50K mid pot in place of the 6K8 mid set R. I'd probably go with 25K. Or put in a 10K and put the 6K8 between the pots ground end and ground. That way with the pot on 0 it will be stock, and as you turn up the mid pot you will add mids.

Very popular mod with guys. I don't think you'll regret it. 

[Lectroid]

PRR, Willabe, slicker,

Thank you all for passing on all of this acquired wisdom, much more than I'd hoped for.  You've cleared up a lot for me about the theory and practice behind filtering and coupling caps.

@PRR,  I copied the p/s from a previous project without updating values--I'm a hurry to get building this thing.  (This amp-building thing is really addictive, isn't it?)  But it was impolite to give you a circuit that was wrong.  I may be a simple hacker now, but I aim to do better.

In the actual Princeton-clone circuit, I now have planned a can capacitor, 30u-30u--one for the reservoir and the other to filter the screen voltage stage.  I had planned to add another 22u for all  four preamp tubes and call it a day.  Now I'll add a second 22u caps and separate out the gain stages, one 12AX7 tube to one node.  I've also dropped values on the coupling caps as you suggested to limit blocking.

@Willabe, sluckey,  thanks for the discussions of Fender amps designed for Fender guitars.  It explains a lot about the different makers' amp designs. When I was in London many years ago, I hit every guitar store I could find and I was struck by how under-represented Fender guitars were in England.  Many many Gibsons and other brands with humbucking pickups, but very few with single coils.  Lots of Marshall amps though, which probably explains it.

@Willabe, good tip on the middle control.  I'm tight for space in a small chassis but now you've got me trying to figure out how to get that one in there.  It will definitely be on my next build.  (Always a next one.).

Also, can you elaborate on how the stages interact with the p/s when there are too many stages on one p/s node?  If there's a brief explanation...  Feel free to point me toward a reference if it would take too long in a short space.  I don't mind reading.

You guys are great!      Thanks again.

[Willabe]

PRR, Willabe, slicker,

1st, it's Sluckey. (Auto check spell?)

In the actual Princeton-clone circuit, I now have planned a can capacitor, 30u-30u--one for the reservoir and the other to filter the screen voltage stage.

You should read this before you put in a dual cap can that shares it's single ground lead and before you wire up the amp.

http://www.valvewizard.co.uk/Grounding.pdf

Also, can you elaborate on how the stages interact with the p/s when there are too many stages on one p/s node?  If there's a brief explanation...  Feel free to point me toward a reference if it would take too long in a short space.  I don't mind reading.

PRR did, look at the 1st/top attachment in reply #5. 

[sluckey]

In the actual Princeton-clone circuit, I now have planned a can capacitor, 30u-30u--one for the reservoir and the other to filter the screen voltage stage.

You should read this before you put in a dual cap can that shares it's single ground lead and before you wire up the amp.

I've had excellent results doing exactly that. He plans to use one cap for the plate node and the other for the screen node, both of which should be grounded to the same power ground. Great idea IMO.

[Willabe]

He should still read it for grounding wiring layout scheme.

[Lectroid]

Sluckey, Willabe, PRR:

Triple-damn that spell-check demon!  (Almost got me fired once...)

Yes I missed PRR's reference to the RDH4. Just got the 4th Ed. a few weeks ago.  It's mammoth, but the explanation he pointed me to is very clear. I now see the reason/need to isolate each preamp tube's p/s node.  Got it.

Second, I've read the grounding section of VW's book a few times and do try to separate the p/s grounds from signal chain grounds.  But I'm still assimilating it, and you all saved me from a major pitfall.  If the amp had oscillated, I don't think I have the tools to have figured it out.  I'd have just dumped it and started over.

Third, I see there's disagreement about my use of the two-can cap here.  I think I'll rely on sluckey's conditional approval for use only on the plate node and screen filter.  Past the screen node, discrete caps on each node will be my best practice going forward.

One last question on this topic is one I have about putting a drain resistor to drain the caps when the amp is switched off.  When I tried installing one like in the drawing, with a power rating of 2W, it got hot enough to begin to smoke.  Do I need a higher wattage resistor or am I wiring it wrong?  Does every cap need it's own resistor?

Gentlemen, thanks again for all your help,  you're the best.

[sluckey]

I'm surprised that your 470K/2W resistor is smoking! Are you sure it is 470K? I've used a lot of 220K 3W metal oxide resistors as drains and never seen any smoke. I don't touch but they don't even look like they have even been hot.

I only use one drain resistor unless I have two series caps.

[Lectroid]

Okay, thanks for validating the correct hookup.  I'll try it again.

[sluckey]

Third, I see there's disagreement about my use of the two-can cap here.  I think I'll rely on sluckey's conditional approval for use only on the plate node and screen filter.  Past the screen node, discrete caps on each node will be my best practice going forward.

Some people think there's only ONE correct way to do grounding. But dual cap cans are very popular. Heck, where would Marshall be without dual 50+50 cans. Their amps are full of them! And what about the popular and successful Princeton Reverb? It has all it's filter caps in a single four section can. I don't see any clones that are using the one and only correct cap grounding scheme.   

[Willabe]

But dual cap cans are very popular. Heck, where would Marshall be without dual 50+50 cans. Their amps are full of them!

Yes and the early Marshalls have always been on the noisy side. (And weren't some of the Marshall dual cap can's wired || as 1 cap?) Certainly more buzz than the old Fenders of the same power and time frame. Most/a lot of amps buzzed to some degree. That's just the way it was back then and if you wanted the Marshall sound, then you put up with it. Players just got used to it. That doesn't mean that should be our standard of signal to noise ratio.

Between songs you turned down the guitars volume control. Even though it wasn't as buzzy as a Marshall, we all did the same with Fenders too.   

Live it's not always as much a problem, audience noise masks a lot of it. In the studio (or recording at home) it can be a problem. At home it can be annoying.    

Some people think there's only ONE correct way to do grounding.

To me it comes down to; how much buzz is too much?   

Any buzz, noise, is too much for me  -IF-   it can be eliminated. I don't want to know my amp is on by hearing buzz. 

I don't think there's only 1 correct way to do grounding, but I do think there are ways to wire grounds that can be taken further and further, 1 at a time  - in steps -  that are pretty much understood and cab be prioritized in importance for how much each step decreases the noise floor level and it adds up.

Like the PT B+ CT going directly to the 1st B+ filter caps negative lead. That is going to give a higher reduction in buzz/noise than other ways of wiring it. 

Can you use a dual cap can there for the B+ power tubes plates and screens? Sure, we've seen it done. Is it as quite as using separate caps? Can we hear a difference, or enough of a difference, between the 2 for it to be worth the change?   

The screen grid is much more sensitive than the plates, so grounding the screen cap with the 1st B+ caps high ripple could modulate the screen cap. The explanations out on the web/books that many of us have read on this point make sense to me, have I tested it on the bench, no. But this would still be a high priority in wiring up the ground for me. It can't hurt, can only make it better.

Edit; Marshall did use a -FB loop and they were PP power amp, noise cancellation through the PT, so maybe between 1, the other, or both it killed off most if not any buzz you could hear caused by a dual cap can in the plate/screen PUP? Might be more of a problem in SE amps and/or amps with no -FB loop around the power stage? And Ken Fisher did recommend lifting the grid wires up off the chassis and leaving them in the air. He said it often worked to reduce noise in old Marshall's. I never tried it, but he worked on dozens of old Marshall's.     

Back then most of us had no idea why our amp(s) were noisy. Now after thousands of guys putting all these amps under the microscope, we have a lot of researched information and answers to go by. And since most of use are not mass producing amps, we can spend a little more $$ and time building them. 

Anything that can make the amp quieter, without changing the tone, and breaking the bank, I'm for.

Wired grounding, in stead of random grounding through the chassis, including B+ filter caps distributed across the chassis. Line noise caps on the secondary and primary of the PT. Also PT with both an internal shield and a copper external shield, if I can get. DC elevated to 70dcv/80dcv with a humdinger pot. Twisted pairs, 3's or more of wires any where I can use them. FW or FWB -bias, if I can get the PT for it, or pull it from the main B+ and dual -bias pots.

All of it.  

[Willabe]

And what about the popular and successful Princeton Reverb? It has all it's filter caps in a single four section can.

I've been thinking about this.       Fender did separate the circuit chassis grounds from the eyelet board and pots/jacks from the power chassis grounds. But the can cap does tie them back together.

I'm wondering if;

1. Because it's a low power amp, 15w, that they got away with it because lower PT B+ current draw = lower ripple B+ cap charging currents?

2. Somehow with the lower current, the ground connections traveling through the chassis back to the filter cap can ground and chassis power ground filter out the ripple/buzz/noise some what? What if each of the preamp/verb/trem/PI grounds were run back with separate wires, a true star ground set up, to the power amp/can cap ground, instead of through the chassis, I'm thinking it would buzz like crazy? 

3. It has a 10" speaker. So like an 8'' speaker in a Champ, it filters out some of the low end buzz?

4. Some combination of the above 3? So a 35w 2 x 6L6GB/C amp with 2 x 12" speakers wouldn't work with that cap can/grounding setup?

[sluckey]

OK, you can be in charge of Ground Control. But leave standby switches alone. pdf64 has that covered.   

[Willabe]

OK, you can be in charge of Ground Control. But leave standby switches alone. pdf64 has that covered.   

          


Uh-oh, I'm for those too! 

[Willabe]

..... And Ken Fisher did recommend lifting the grid wires up off the chassis and leaving them in the air. He said it often worked to reduce noise in old Marshall's. I never tried it, but he worked on dozens of old Marshall's.

This would seem to back Fisher, from R. G. Keen;

"Wires near the chassis are stitched down to the chassis by capacitive connections to it like Gulliver in Lilliput, and this connection helps eat the capacitive radiation from the wire. Chassis are usually steel, and high current wires near the chassis have much of their magnetic radiation shorted by the ferromagnetic steel. High impedance wires like input grid wires can be loaded down by the parasitic capacitance of running right next to a chassis." 

Here's the page it's from;

http://www.geofex.com/Article_Folders/lead_dress/lead_dress_in_tube_amps.htm

[PRR]

But dual cap cans are very popular. ...

....Between songs you turned down the guitars volume control. ...........................

Guitar hum is not amplifier buzz.