Higher gain out of a tube hi-fi reverb amp

Recording Techniques, People Skills, Gear, Recording Spaces, Computers, and DIY

Moderators: drumsound, tomb

Post Reply
atavacron
steve albini likes it
Posts: 348
Joined: Mon Nov 03, 2008 2:50 am
Location: Oakland, California
Contact:

Higher gain out of a tube hi-fi reverb amp

Post by atavacron » Tue Jun 21, 2011 8:19 pm

I'm modifying a Pioneer SR-101 stereo reverb in order to use it at professional balanced line level. It's not a bad piece of consumer hi-fi kit once you mess with it a bunch. Here's the schematic:

Image

I swapped the channel-crossover stuff happening between the cathode of the first triode of each channel and the opposite channel's last triode's cathode. I also swapped the inputs. I also separated the ganged 500k reverb mix pots into two discrete 1M log pots, as that's what I had lying around. In other words, the unit is now dual mono. 'Lytics and balancing transformers on the way - got two 15k:15k for the ins, and two 15k:600 for the outs.

Here's the question: I need to make up about 12dB more gain in the output section in order to avoid having to use a line amp after this guy. Can someone tell me which values to change around the 12ax7 of each channel in order to squeeze out some more juice? I have a feeling this has something to do with the plate resistors and the B+ voltage coming into them...or do I just need to add cathode bypass caps on the last two stages? Advice plz! I'm not so good with the math here.
shipwreckoakland.com
5 1 0 . 9 7 2 . T A P E

lacquer_monkey
gettin' sounds
Posts: 100
Joined: Sun Dec 25, 2005 7:17 pm

Post by lacquer_monkey » Thu Jun 23, 2011 3:19 pm

The quickest way to add some gain would be to add bypass caps to the cathode resistors of the recovery and output stages. 25uF/25v should work fine.
uh... what??

atavacron
steve albini likes it
Posts: 348
Joined: Mon Nov 03, 2008 2:50 am
Location: Oakland, California
Contact:

Post by atavacron » Thu Jun 23, 2011 11:17 pm

Thank you! Did it. Used 100mf caps 'cause that's what I had lying around.

Lowered the 100k plate resistors on all triodes to 56k. Changed the 22k feedback resistors in the driver stage to 100k.

Got rid of the "Dry" RC feeds from the last triodes to the first triodes, since I only need the wet signal. Kept just a 2.2k as the input triode's cathode resistor.

Channel 1 sounds sweet, getting a 6db boost on the outputs at max volume, marked a unity gain notch for each channel on the faceplate at 3 o'clockish.

Channel 2's tank is bad, gonna use a similar accutronics with an 8ohm in and a 2250ohm out...added a 5.5k:8ohm 5W output tranny to my Edcor order to drive it with the 6BM8.

Fun!
shipwreckoakland.com
5 1 0 . 9 7 2 . T A P E

lacquer_monkey
gettin' sounds
Posts: 100
Joined: Sun Dec 25, 2005 7:17 pm

Post by lacquer_monkey » Mon Jun 27, 2011 10:29 am

Lowering the value of the plate resistors will actually slightly reduce the gain of a tube stage. Gain is dependent largely on the ratio of cathode circuit resistance to plate circuit resistance. There are a lot of factors involved (tranconductance, plate resistance at the operating point, resitance of next stage input) besides just the resitors. But in general, lowering the plate resistor by itself will lower gain.
uh... what??

User avatar
Scodiddly
genitals didn't survive the freeze
Posts: 3958
Joined: Wed Dec 10, 2003 6:38 am
Location: Mundelein, IL, USA
Contact:

Post by Scodiddly » Mon Jun 27, 2011 4:19 pm

That's the thing about plate load resistors - they have to pass some DC current to the plate, but since the PSU end is at a virtual AC ground (ie all those filter caps - any AC is shunted to ground), the lower the resistor value the more AC from the tube is leached away.

atavacron
steve albini likes it
Posts: 348
Joined: Mon Nov 03, 2008 2:50 am
Location: Oakland, California
Contact:

Post by atavacron » Tue Jun 28, 2011 6:55 pm

...or in the case of a stereo amp, the AC can cross over to the other channel! There was definitely a point in this process when I was getting signal from one channel on the other and vice versa.

Thanks for the education about gain ratio guys! This is great. Makes sense too, as my gain dropped 3dB or so when I lowered the plate resistor values, but jumped up 6-9dB when I added the cathode bypass caps. Don't quote me on those numbers.

What also happened when I increased the voltage to the plate was a significant increase in headroom...which is initially why I swapped out the plate resistors. Is that accurate, that a higher plate voltage generally offers the opportunity for better headroom?
shipwreckoakland.com
5 1 0 . 9 7 2 . T A P E

User avatar
Scodiddly
genitals didn't survive the freeze
Posts: 3958
Joined: Wed Dec 10, 2003 6:38 am
Location: Mundelein, IL, USA
Contact:

Post by Scodiddly » Tue Jun 28, 2011 7:35 pm

Higher voltage will of course give more headroom - that's the same thing as the "rail voltage" in solid state circuitry. Of course the higher voltage also means that there will be higher current flowing through the various loads, so you need to take that into account by maybe increasing some resistor values.

atavacron
steve albini likes it
Posts: 348
Joined: Mon Nov 03, 2008 2:50 am
Location: Oakland, California
Contact:

Post by atavacron » Tue Jun 28, 2011 8:13 pm

So if you have a 1k cathode resistor and a 100k plate resistor on a triode with a mu of 100, does that give you a gain of 100:1 in theory, provided the DC plate voltage is higher than 100x the AC input voltage?
shipwreckoakland.com
5 1 0 . 9 7 2 . T A P E

lacquer_monkey
gettin' sounds
Posts: 100
Joined: Sun Dec 25, 2005 7:17 pm

Post by lacquer_monkey » Tue Jun 28, 2011 11:01 pm

Not exactly... The plate resistance of the tube, the plate resistor, and the grid resistor of the next stage are effectively in paralell. This network is the plate circuit resistance. The cathode circuit consists of the cathode resistor in series with the cathode resistance which is basically the reciprocal of the transconductance at the operating point. When you bypass the cathode resistor, the cathode circuit resistance becomes more or less the reciprocal of the transconductance at the operating point. That's where the jump in gain comes from.
uh... what??

lacquer_monkey
gettin' sounds
Posts: 100
Joined: Sun Dec 25, 2005 7:17 pm

Post by lacquer_monkey » Tue Jun 28, 2011 11:03 pm

This is all interms of AC (signal) rather than static DC stuff...
uh... what??

atavacron
steve albini likes it
Posts: 348
Joined: Mon Nov 03, 2008 2:50 am
Location: Oakland, California
Contact:

Post by atavacron » Wed Jun 29, 2011 12:56 am

I think I need to do a lot more reading...

Hey why don't they have classes in this any more? I would absolutely go take a night class at a local tech college if I could find one where they were teaching tubes!
shipwreckoakland.com
5 1 0 . 9 7 2 . T A P E

lacquer_monkey
gettin' sounds
Posts: 100
Joined: Sun Dec 25, 2005 7:17 pm

Post by lacquer_monkey » Wed Jun 29, 2011 12:43 pm

If you mean classes in tube electronics, it's because nobody really uses tubes anymore except in certain niches like big radio transmitters and RADAR installations. Almost all community college classes have classes in math for electronics and AC and DC circuit analysis. Also, N-channel JFETs behave almost exactly like a tube so studying that can illuminate some things. There are books available in local and college libraries that cover all this stuff. A lot of the places that sell supplies for tube amps also sell books about tube amp theory.
uh... what??

themagicmanmdt
george martin
Posts: 1347
Joined: Mon Nov 13, 2006 8:47 pm
Location: home on the range

Post by themagicmanmdt » Wed Jul 27, 2011 10:55 am

you're going about this all wrong, guys.

the reason why you don't have enough gain into a line input is because your outputs are high impedance and are wanting to be terminated as such.

feeding the plate from v3 - the load is looking for upwards of 470k (since it's a 12ax7).

you're losing all of your signal due to improper impedance termination!


boosting gain inside of the amp might help it out a bit, but you're going to start monkeying with the designed gain staging and headroom within the amp by bypassing everything for more gain. plus, lowering your plate resistors will decrease gain a little (which is good to make up for the bypassed cathode, which somewhat dramatically increases gain) but also changes your bias point....

by using a 15k:600 output transformer, make sure you terminate into 10k or higher line inputs. make sure the OT can handle a super healthy level (+20dbu or higher) because you're going to drive it harder to get a good line output after the OT steps down the voltage...

impedance ratio is 25:1; turns ratio is 5:1 on a 15k:600.

a 10k load reflected to the tube will be 250k....which to me is the minimum to live with for a 12AX7 to be terminated into for good sonics and low signal loss.

i'd say hold off on anything further as far as all these mods. you should be OK, but you might hear odd distortion or some unclean (maybe in a good way? maybe in a bad way?) stuff going on with the changes you made. you may end up putting things back to stock...or at least working with your bias points again.

try the OT's first. make sure the output .01 caps are replaced and fresh (get sprague orange drops, they sound great!) because you don't want any DC on your OT's.

the last option would be to have a transformerless cathode follower stage built after the output gain stages. you still need to terminate into at least 10k, but you skip the transformer. some like the sound, some don't. i'm 50/50 depending on what i want to hear out of the gear.



groovy

ps - replace all caps - everything sounds better with new caps all around - not just electrolytics - things that are 100% 'stock' just sound different.... not 'better'.

tell us what you find and how it sounds!
we are the village green
preservation society
god bless +6 tape
valves and serviceability

*chief tech and R&D shaman at shadow hills industries*

atavacron
steve albini likes it
Posts: 348
Joined: Mon Nov 03, 2008 2:50 am
Location: Oakland, California
Contact:

Build update

Post by atavacron » Thu Mar 15, 2012 6:16 pm

So here's the completed build...http://www.flickr.com/photos/shipwrecks ... 5289/show/. The internals weren't quite done when I had it splayed open for photos, but everything is now connected and buttoned up.

I used an Edcor GSXE5-8-5.5k single ended output transformer for the 6BM8 pentode on the right channel as the reverb driver, with a 220Ω screen resistor from the screen tap. This transformer is total overkill for a ~1W output, but that's what I could find for cheap. I had to mount it in the lid of the unit in between the balancing transformer circuit boards, so there is something like 14" of shielded 22AWG 3-core cable running from the pentode to the transformer and back for B+, Plate and Screen. I dabbed some liquid electrical tape on the contacts...looks like there's only about 100VDC on those anyways. I got rid of the 4.7kΩ and 47kΩ plate and screen resistors, as well as the 3?F/150V cap from plate to ground. It basically looks like any single ended output stage now.

I also star grounded the thing and shielded all internal signal cables. Should be pretty quiet.

I'll do some listening tests and post them. Hope nothing explodes.

Thanks themagicmandt for the reality check...
shipwreckoakland.com
5 1 0 . 9 7 2 . T A P E

themagicmanmdt
george martin
Posts: 1347
Joined: Mon Nov 13, 2006 8:47 pm
Location: home on the range

Post by themagicmanmdt » Tue Apr 10, 2012 2:09 pm

whoa. i was wayyyy too passionate in my previous post. i even underlined something?

jeez. i gotta lay off the caffeine/sugar/tubes.

glad it sounds good.
we are the village green
preservation society
god bless +6 tape
valves and serviceability

*chief tech and R&D shaman at shadow hills industries*

Post Reply

Who is online

Users browsing this forum: No registered users and 39 guests