Simple SE guitar amp: Tuning the 'gain'.....

[Wharfcreek]

I've built a little Single Ended guitar amp. Pretty simple thing; one 6BQ5 for the output, with one 12AX7 as a driver. Given the two triodes of the 12AX, I actually have what I guess you'd call a 'series' driver, as the first section is feeding the section which drives the output tube. This is basically a 'spare parts' piece, with a little SE output transformer I pulled from an old 'Hi-Fi' amp made by 'US Audio'. I actually believe this to be an Asian company using the name of 'US Audio' as a USA Marketing ploy. None the less, it came to me with only one good OPT, so I swapped them both out and had this as surplus. I also had an old OPT that I was able to get a HV of about 250V out of after setting up the output section. I don't recall all the details, but I've got a 150K resistor with 100uf / 100V bypass cap, and as I recall I think the bias calculated out to be around 35ma on the output tube. So, moving on to the 12AX, I think I'm running around 120VDC or so on the plates of both triode sections. As, as per what I consider 'typical' Guitar amp 12AX cathode set-up, there's a 1.5K resistor to ground on each triode cathode, along with a 25uf/25V bypass cap.

So, now comes the question: Can I change the values of the cathode resistors on the 12AX section in order to 'adjust' the gain going to the output tube?.....and, if so, is 'more' resistance equal to 'less' gain? I guess I'm just thinking about output tubes here....where increasing a cathode resistor value would result in less current draw by the output tube....and reduce output power......or so I believe. I may be wrong....and if so, PLEASE correct me. Anyway...as my amp stands now, using the 12AX as the 'driver' tube.....it's just too 'gainy'! I've played around with other tubes, including a 12AT, a 12AY, a 12AV, and a 12AU. I found this substitution to be very revealing, as the 12AX provided way too much early gain, where the 12AU on the other end of the scale provided just enough gain to drive the amp to a very minimal distortion level. The AT and AY seem to be the sweet-spot for the amp.....but I'm simply wondering if I can go back to the 12AX and tame the amp through the circuit vs changing tubes? Perhaps the answer is to drop the voltage to the plates? Just looking for some advice here from someone who's maybe played around with this and has some experience with these kinds of issues.

Many thanks,

Tom D.

 

[kward]

There are several ways to reduce the gain of a common cathode gain stage:
1. Decrease the plate resistor value
2. Increase the cathode resistor value
3. Remove the cathode bypass capacitor, it if has one

These first two things affect bias, so you'll have to work out how those changes affect bias and compensate (such as decrease the rail voltage), or at least be aware of how the bias changes when you change resistor values to make sure the stage is still operating within its safe area and will still give you the voltage swings you need.

The other method you could maybe try, if the first stage is capacitor coupled to the second stage, is to put in a pot after the coupling cap but before the grid of the second 12AX7. This would give you a variable voltage attenuation network, just like a volume pot would do, although you may want to use a linear pot in this case.

Disclaimer: No guitar amp experience here, just thinking through the theory.

 

[Wharfcreek]

K, thanks for the commentary. Now let me see if I understand it correctly:

#1) In 'decreasing' the Plate Resistor value, will that cause an 'increase' in Plate voltage? If so, I guess I'm confused about how a 'higher' voltage would actually result in 'less' gain from the tube.

#2) Increasing the cathode resistor value: that one I think I get......as I stated above. The 'typical' values I see on Fender amps as well as others is 1.5K with the 25uf/25V bypass cap. So, to 'decrease' gain, I would perhaps move that resistor value from the 1.5K to maybe something in the 2.2K to 2.8K range....just as a 'for-instance'???

3) Removal of the bypass cap: Had somewhat considered that as well. Some of the 'early' Fender designs of smaller SE amps, using both a 12AX7 and a 6V6 output tube, some of them actually had either 1 or 'none' for bypass caps on the 1.5K resistors. Still 1.5K though. I had wondered about removing those, but wasn't sure how that would affect the frequency range of the amp. One other thing I failed to mention is that the amp sounds rather 'dark' to me......like the HF may be rolled off. Or, this may be a function of the loss of 'harmonics' to some degree......those 'second' or 'third' order harmonics I hear tell about......... Frankly, this is all still over my head. But, at least I'm thinking about it.

Anyway.....I think I'll try the removal of those bypass caps and just see what it sounds like. I might add that the amp's volume control actually is the 'pot' you referenced above. There is a 1Meg pot that comes after a 1st to 2nd stage coupling cap. I believe the 'tone' control is in there ahead of this VC as well, though I don't have the amp or schematic in front of me at the moment.

One thing I have to say is that I believe all this is going to help considerably in my overall understanding of 'tube' amps! I had played around with guitar amps some 20 odd years ago....but never really doing any more than just following a schematic and looking for something either broken, disconnected, shorted, or otherwise 'blown up'!!...lol. While I did 'construct' a few amps......again, it was just following the diagrams using parts that were basically 'replacement' values. I never really had to 'figure anything out'....so to speak. This little project has been somewhat similar in that I started with a couple of 'known' schematics. But, in the end, there's really nothing that actually follows these other diagrams outside of the basic 'conceptual' stuff. Yes, it's a typical little cathode bias SE EL84 amp. But, in having started with an unknown power transformer, I've had to do a lot of back-pedaling in order to even get the amp working. Now, this 'fine tuning' stuff......it's helping me have a much better understanding of how each stage works, as well as how the components affect each stage. All good stuff!!

Tom

 

[s-petersen]

You can always use a little negative feedback to reduce gain.

 

[billyz]

I find it easiest to change the cathode resistor in the v1, then v1b. Are the cathodes sharing the cathode resistor?
Usually the lowest I go is 820 ohm to each cathode for more gain. Or the lower gain use 2.7k each. Bypassing increases gain and can tune the frequency balance. 25uf is considered full range for guitar. 0.68uf will give a tilt to the higher frequencies. I have used 0.47 as well.
The plate load resistor can also raise or lower the gain , with 100k being pretty average. 220k is higher and crunchier, I have used 82k for some jazz players .

You can add treble with a treble peaking resistor and paralleled capacitor. Try a 470k resistor paralleled with a 470pf capacitor. This would be in series between the v1a, v1b. These are just examples . Another method to increase treble is to bypass the volume pot with a 250pf cap . Other values can be 120pf to 500pf to taste.

Sometimes it’s not adding treble but reducing the lows that is needed. For example use a .002 coupling cap instead of .02uf. I have even used 500pf coupling cap to trim the low end .

 

[Wharfcreek]

Scott, I was just looking at a Schematic of an old Fender 'tweed' Champ....... and that uses some negative feedback which, if I recall, was Leo Fender's answer to attempting to 'clean up' the sound of some of his amplifier. So, I may give that a shot as well.

Billy, I like your ideas as well. The nice thing about this little amp is that it's built in such a way that I can pretty easily attempt any or all these suggestions and 'learn' from them. To that end, I'd like to just clarify your 'treble peaking' r/c suggestion. When you say 'in series between v1a and v1b...... does this mean that I leave the rest of the circuit in tact, but just plug this in along side of it? These are the 'cathodes' of the 12AX that you're speaking of, yes? So while leaving the cathode resistors in place, I just put this little r/c network between pins 3 and 8, yes? Sorry if I'm not reading that as you intended......but that's how I'm understanding it right now. I'd appreciate a clarification.

As to the 'coupling cap' part of this, I also want to clarify that you're referring to the cap between the 12AX and 6BQ5? I only ask because the two triodes of the 12AX are also coupled with a .02uf as well as the 12AX to the EL84.

One thing I also ought to mention is that I 'assembled' this amp by modeling both a Fender 5E2 schematic, as well as the Antique Electronic Supply little 'kit' version of an old Tweed amp.....that AES refers to as a MOD 102. I wanted only a 1-knob tone control, and the 5E2 provided that in it's front end. So, I used that circuit design for that part of the amp build. But, on the Fender amp the plates of the 12AX are listed as having 150VDC....where as the MOD kit has a plate voltage on V1A as being 121VDC and V1B as being 115 VDC. I think I 'compromised' and set both at about 120 to 130. Maybe tonight when I get home from work I'll try to draw out my build 'as is' and post my own schematic.

Tom

 

[kward]

#1) In 'decreasing' the Plate Resistor value, will that cause an 'increase' in Plate voltage? If so, I guess I'm confused about how a 'higher' voltage would actually result in 'less' gain from the tube.

Yes, decreasing the plate resistor value while keeping the rail voltage the same will in fact increase the voltage seen at the plate (all else being equal). It also increases the quiescent current through the tube, but the overall effect is to lower gain. The most accurate short answer to explain why this is true is that decreasing the plate resistor increases the slope of the load line, and a steeper slope on the load line means less output voltage swing is available for a given input voltage swing, thus lower gain. The most gain a common cathode gain stage can ever have under ideal (perfect) conditions is the mu factor of the tube. For a 12AX7, that being a value of 100. So lowering the plate resistor has this effect of chewing away more at the available gain.

Incidentally, the opposite is also true--increasing the plate resistor value will increase the gain of the stage, everything else being equal. This is why folks sometimes use constant current sources instead of plate resistors for plate loads. It's because a constant current source looks like a very very large plate resistor to the tube under dynamic conditions (at the 1 to 10 Megahom value for a good CCS). This has the effect of keeping current constant, or in other words flattening (reducing slope) of the load line, which provides more output voltage swing available for a given input voltage swing, thus higher gain.

 

[Hi-FiGuy]

For giggles you should try an ECC832 or ECC823 those tubes have one half of a 12AX7 and 12AU7. The difference between the two is which triode comes first in the circuit. The beauty is that no rewiring is necessary. You just plug an play.

 

[s-petersen]

7247 12dw7? also

 

[Wharfcreek]

K, not to belabor this, but in trying to relate what you have just posted above to what Billy has posted in #5 above, seems to me I'm reading these two responses as being directly contrary to each other. Billy mentions going from the 1.5 to an 820 to 'raise' gain. But, if I'm reading your post correctly, what you're saying is that a reduction in cathode resistor value (ie going from the 1.5K to 820 ohms) would actually result in a 'reduction' in gain. So, I'm wondering if the problem here might be 'language' or terminology? In any event, I think it's up to me to simply 'try' some of these suggestions and see what kind of result I actually get in terms of performance. I find myself somewhat at odds with myself in that I've actually 'almost' solved the issue just by swapping the 12AX for a 12AY. That tube seems to deliver the best level of overall drive in the amp such that the unit can be turned down and still have some level of volume while playing 'cleanly'....and when turned up the drive clearly peaks and things start to 'over-drive' such that the amp 'dirties-up' without going too far and just sounding horrible! An example of this would be what some of the solid-state guitar amps sounded like when they were given 'overdrive'........where they just saturated so badly that virtually all character and tone was lost. While not a good one, I'm a mediocre 'blues' guitar player......and while 'country pickin' seems to have the 'clean' sound all wrapped up.....a slightly 'dirty' sound is desirable in blues guitar playing. Not to the level of the Metal heads......but somewhere between 'country' and 'hard-rock'. This little SE project is me 'learning trip' in trying to duplicate that sound. A lot to learn......and a LOT different than aspiring for a 20 to 20K 'no-distortion' Hi-Fi amp!~!! Tom

 

[Wharfcreek]

If I PM you guys, can you send me a couple of those tubes to try??? Willing to pay...... I recall the 7247 as being the driver tube in the Dynaco ST-35. I recently did a project of building that amp only using Shannon Parks' driver modification of one 12AX7 and one 12AU7. That amp build has been in my living room ever since......and sounds great! I may have one more of those but I'm not familiar with the ECC832 or ECC823, but it sounds like something I should have to try!! TD

 

[s-petersen]

I would have to look for the 7247's I have, I don't have many

 

[kward]

Tom, reducing the cathode resistor value will increase gain (if it is not bypassed). Reducing the plate resistor value will decrease gain.

 

[mike567]

try using a 1uf cathode bypass cap on one or both 12ax7 sections in place of the 25's

 

[billyz]

To that end, I'd like to just clarify your 'treble peaking' r/c suggestion. When you say 'in series between v1a and v1b...... does this mean that I leave the rest of the circuit in tact, but just plug this in along side of it?

Not along side as in parallel. I mean in series with the signal path.

I edited a well known guitar amp that includes all these ways to increase treble response.

 

[primosounds]

try using a 1uf cathode bypass cap on one or both 12ax7 sections in place of the 25's

So reducing the size of the cathode bypass cap also effects the gain. I was wondering if the relationship of gain / the cap capicitance is proportional?

 

[billyz]

I do think some of the confusion is terminalogy.
1. cathode resistor, 820 has more gain than 1.5k, 1.5k has more gain than 2.7k.
so to lower gain with the cathode resistor go from 1.5k to 2.7k ( try 3.3k ).
2. cathode bypass capacitor, 330uf increases gain at all frequencies, 25uf (20, 22uf) is conssidered full range for guitar which has a low frequency of about 80hz maybe 60hz.
1uf or 0.68uf will increase gain of only the higher frequencies ( the exact knee depends on the circuit impedance etc).
3. Plate load resistor, 100k has less gain than 220k, 330k has more gain than 220k, 82k is less gain than 100k.

you can mix and match these values, but dont go crazy because you really should do a load line to see where you are at, and the plate voltage and tube type are factored .

for a typical 12ax7 stage, to lower gain I prefer to use a 12ay7 , a 5751 is a little less than a 12ax7,
you can also use others , 12at7, 12au7, 12dw7( different triode sections ).

 

[billyz]

Having FUN

 

[Hi-FiGuy]

The 12DW7 / 7247 is the same as the ecc832. I believe using that tube the 12au7 section would be first triode. Ideally you would want the ax7 triode to be first. Then the au7 section would be the driver, thats where the ecc823 comes in. Would be fun to try this unequal dual triode, give you a unique sound. If you wired the triode sections backward with respect to the fender amp schematics the 12dw7 would be ax7 then au7. If you drop in an ax7 tube both sides being equal you could modify the feedback network to drop gain as well.

 

[Wharfcreek]

Billy,

I'm at work and haven't had much time to study your latest contribution (thank you very much). But, in looking at it, I see what appears to be the 'bright' channel....the one with the 470K and 500pf coupling it to the next gain stage. HOWEVER, in looking at that section of the 12AX, it appears it's cathode resistor and Bypass cap are 2.7K and 0.68 respectively. So, that being true, is there some kind of 'compensation' going on where the gain between the two channels becomes roughly equal.....or does the 'bright' channel also carry more gain but does so simply at higher frequencies?