Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Systemloc on February 19, 2011, 04:31:35 pm
-
Hello!
I'm aiming to build something loosely based on a Blackface Champ using tubes from my junk pile. I'd really like reverb in it, so I've been comparing the Blackface Champ, Princeton, and Princeton Reverb. All three appear to use the same first two pre-amp stages, and the Champ has a simple class A power stage rather than a phase splitter and push-pull power stage.
The reverb is a whole other issue. There is a parallel 12AT reverb driver and 12AX7 recovery, but there's also that additional gain stage in the main signal path! I've read that the additional gain gives the PR more drive and changes the tone a bit. Notably, in the input to that extra stage, there's a resistor divider on both the main signal path and the reverb path. Obviously, there's too much gain in three 12AX7 stages on the main path, so the divider cuts about 1/4th in the main path (1M:3M).
Estimating overall gain into the splitter in the PR, roughly 30*(1/10)*40*40*(1/4)=1,200.
Using the 12AX7 graphs and the 100k Rp load line, I estimate mu=70. The load of the tone stack will lower that first stage gain further to 30. The tone stack also has about a -20db drop, which is the 1/10 gain. On the other two stages, a 1M load lowers gain to 40. Finally, the 1:3M divider cuts the signal by 1/4th.
Estimating overall gain into the splitter in the Princeton w/o reverb, 30*(1/10)*40= 120.
Same as above, minus the resistor network and extra stage.
Holy cow!! That extra stage gives 10x more gain! Of course, this is open loop. I haven't worked out how to calculate how much negative feedback there is to calculate closed loop gain. That may even things out.
---
Now, building a junkbox Champ. Yeah, I can buy the parts, but I like the challenge and learning that goes with using only what I have on hand.. I don't have 12AX7's, so I'm evaluating different triodes.
One note I came across:
By the way, the 12AT7 in a first gain stage is an awful tone generator in a guitar amp. If you want to drop front end gain use a 5751 (gain of 60-70) or a 12AY7 (gain of 40). The 12AY7 was the first gain stage in the classic Fender Tweed Bassman, Deluxe, etc."
http://www.guitaramplifierblueprinting.com/files/Phaseinverter.pdf (http://www.guitaramplifierblueprinting.com/files/Phaseinverter.pdf)
I wonder if this is due to the plate resistance. The tone stack is a heavy load, and will drag gain down unless that plate resistance is low. What the above quote suggests is that you want a heavily loaded first stage.
I'm also thinking with the difference in gain between the Princeton with and without reverb, I don't need to worry about achieving as much gain as with the 3 12AX7 stages, and if I need more, I can use an extra pot to tweak the 3:1 resistor divider into the third stage and get more gain back there.
Main signal path:
So, initially, I was thinking of using a 6MN8 (Triple triode, 47mu, 6250 rp). The low rp would preserve the gain with the heavy tone stack load, thus offsetting the lower gain. However, the above quote put me off of this. Instead, I have two 6SQ7's (Singlet triode 100mu, 91K-rp), and one 6SF7(Singlet triode 100mu, 66K-rp). I'd probably put the 6SF7 as the first stage, as the lower rp better matches that 12AX7 into the tone stack.
Reverb path:
The recovery stage is pretty straightforward, so I'm focusing on the driver stage, a paralleled 12AT7 Pulling a 23K:8 ohm transformer, so a 23K load. Each triode sees half the current, so each is actually seeing twice that load, 46K. Even if only one triode is used, it's well within power limits, so I imagine the point is that paralleling gives you a higher impedance load, which, I think, would be equivalent to a lower plate resistance. The RCA Radiotron Designer's handbook states that the optimal plate resistance for transformer coupling is 6-10K. Higher than this requires more inductance in the primary. The 12AT7's rp is 15K, which, cut in half is 7.5K. That's right in the range specified.
Thus, I need about 60 mu with an rp between 6-10k. I could probably get away with a lower mu tube, and pull more gain from the resistor divider after the recovery stage. I've got a 6BQ7 (Twin triode 37mu, 5.9k-rp), or a 6BC8A (Twin triode 35mu, 5.3k-rp). I could use a single side of either of those. The other side, I might try out as a cathode follower between the first pre-amp and the tone stack. That would take the load off of the first stage and give me more gain. No idea what it would do to the tone. Any thoughts?
Output stage:
I have a couple 6V6GT's. My output transformer is out of an AM radio with a class A 6K6GT. That 6K6GT looks like it likes a higher output impedance, so I may try both and see which sounds better. I've got a bunch of different tubes I can throw in as finals: 6AQ5A, 6GC5, 6JB5, 6AR5, 6CU6.
-
Confessing that I didn't read through all the math, etc., there was a discussion here within the past 6 months about adding reverb to a Tweed Princeton circuit (5F2-A). Tubenit participated - he is very experienced with one-tube reverb circuits - but I can't remember who was building the amp (physconoodler maybe?)
I think that there are two threads, one here and one in the schematics section.
Just offering this up on the theory that starting with a known-good design makes life easier...
Chip
-
I think you mean this thread:
http://www.el34world.com/Forum/index.php?topic=9819.0 (http://www.el34world.com/Forum/index.php?topic=9819.0)
which also references this Champ with reverb:
http://www.el34world.com/Forum/index.php?topic=630.0 (http://www.el34world.com/Forum/index.php?topic=630.0)
Both of these are very informative. A big question I have is that one of these schematics has that extra gain stage in the main signal path, while one does not. I think that extra gain stage, along with the addition of negative feedback, goes a long way towards the difference in Tweed vs Blackface sound. I don't really know, though.
-
Both the Blackface Champ and the Tweed Princeton (5F2-A) have negative feedback. Both have the same number of gain stages (2). The biggest differences are the tone stack and whether or not the cathode on the second gain stage is bypassed.
The negative feedback ratio is different between the Tweed & Blackface circuits:
Tweed = 1,500 / (1,500 + 22,000) = 6.4%
BF = 47 / (47 + 2,700) = 1.7%
That always seems odd to me. I'd assumed that the Blackface NFB ratio would be higher, but it's not.
I think the big offset is that the TMB tonestack of the Blackface Champ has much greater insertion loss than the simple treble cut of the Tweed Princeton. The BF circuit uses cathode bypass on the second gain stage to make up for the tone stack loss.
When you start looking at push-pull Blackface reverb amps with the "extra" gain stage, there's just so much else going on. For starters, that 3.3meg mixing resistor knocks the "dry" signal WAY down. You need an extra gain stage to get the signal back up. Notice that physconoodler ended up with a much lower mixing resistor value. But there's also a phase inverter, and a fixed bias, push-pull power amp with negative feedback seems like a cleaner sounding topology than cathode biased, single ended no matter what else you do. Remember the Tweed Deluxe is cathode biased with NO negative feedback.
I can tell you from personal experiments that lower power rail voltages also contribute a lot to the Tweed sound. For example, drop the plate voltage of the first tube in an AB763 by 100 volts down to 160ish vs. 260ish and you'll get a much warmer, Tweedier tone. Also, going from 350 down to 300ish on the 6V6 plate in a 5F2-A circuit does a lot to get a Tweed sound. Obviously, you aim for higher voltages if Blackface is where you want to be.
Hope that helps,
Chip
-
Hi Fresh Start, thanks for the reply.
I'm really interested in the difference between Blackface Princeton with and without reverb. Like an AA1164, and an AA964. The AA1164 has that extra stage. I've read that even with the resistor divider dumping signal, the AA1164/Reverb has a lot more gain, thus more breakup.
I understand the resistor dividers in the NFB loop, but I'm not sure how to handle the transformer. I would think you'd have to divide the signal level by the voltage ratio of the transformer as well. Do you how that works?
Thanks a ton for mentioning the voltage levels. It hadn't occured to me. I think I'm aiming for a blackface sound. I'm still fuzzy as to the differences. I know the blackfaces break up more to give more of a rock tone. I'll have to play with it :).
-
All other things being equal, lower voltages generate warmer tone and faster breakup. Higher voltages give you more clean headroom. (generalizations but useful ones)
"Rock tone" covers too much spectrum. I've built a Tweed Princeton and a BF Princeton Reverb. The Tweed breaks up faster and has a dirtier tone (in a nice, warm way). The BF PR has beautiful "clean" tones. However, the BF Champ tone probably is closer to the Tweed Princeton given its topology (single-ended 6V6).
AFAIK most guys prefer the AA1164 circuit. The AA964 just doesn't have the umph/balls/whatever unless you just want the clean Blackface tone.
Please don't be offended, but you may need to play and/or listen to different amps to get a better feel for what circuit produces which tone. Studying circuits can only get you so far. You can't go wrong building a BF Champ or Tweed Princeton IMHO as a first amp. I've re-built my 5F2-A twice now and expect to do it at least one more time to get it "right" for me & my guitars.
Respectfully,
Chip