Master Volume on SE Amp - How is to be done ?

[kagliostro]

If you do a search for Master Volume you'll find some versions of it suited for PP amps

but if you do a search about Master Volume specific to SE Amps, there is not a large documentation about

looking on the web you can find a certain amount of SE schematics that has a Master Volume

but if you look to it in detail, you discover that they are poorly engineered

Following one of the schematic that I've seen on the web I initially have think to put a pot following the preceding gain stage and feed

the power tube via the wiper, but this is prone to give problems if the wiper disconnects and DummyLoad pointed my attention to this

and to the fact that the tube (807 - 1625) requires a smaller leak resistor (470k max)



Then I've think to reverse the pot and connect the wiper to the preceding gain stage instead that to the power tube grid, this way if

the wiper fails the powert tube has anyway a grid leak resistor



Thinking at the same problem and searching on the web I've seen a few of schematics (two or three) that uses a pot whose wiper connects

to the power tube grid via a decoupling cap and the power tube has a grid leak resistor (and a stopper), this solution seems the best to me

(note that the value of the decoupling cap is doubled to compensate the fact that they are in series and this halves the capacitance)



What do you think of this method ? Is it a good way to have a Master Volume control on an SE Amp ?

There are better methods to do it ?

MANY THANKS

Franco

[sluckey]

I don't like version B or C. I would use Version A but add a protection resistor from the wiper to ground. If your output tube normally has a 470K grid return then use a 1M pot and 1M resistor. If your output tube normally has a 220K grid return then use a 500K pot and a 470K resistor.

[kagliostro]

Ciao Steve, Grazie

Assuming the tube accepts max 470k as grid leak resistor, if I use a 1M pot and 1M grid leak resistor and the pot is rotated for the max signal

I've 1M grid leak resistor in parallel with 1M pot so 500K equivalent resistor

If I rotate the pot half way (500k) I've 500k in parallel with 1M leak resistor and obtain 333K resistor (fine for the grid of the tube)

but what happens when I rotate the pot to ground ? 0ohm in parallel with 500k will give a leak resistor of 0ohm

(and this idependently of the presence of a protection resistor connected to the wiper and to ground)

isn't this unwanted ?

Thanks

Franco


p.s.: effectively actually solution B has the problem that the load of the preceding gain stage isn't constant and this I can suppose can give problems

[sluckey]

isn't this unwanted ?

No. That's what you want. The grid return resistor value is a maximum value that should not be exceeded. There is no problem with using a lower value, even as low as zero ohms.

The resistor between wiper and ground provides a grid return path in the event the pot wiper fails. True, if that happens, the grid resistor will exceed the maximum recommended value. But that will not harm the tube. But if the pot wiper fails without that protection resistor, the grid return path become an open circuit and that will likely cause the bias to run wild.

[kagliostro]

I understand, thanks



Franco


p.s.: Am I wrong if I think that the 2.2M resistor is useful only if there is a fixed bias but not so much if there is cathode bias ?

[labb]

Take a look at this one. Been several of these built. http://www.ax84.com/static/p1x/AX84_P1x_101004.pdf

[kagliostro]

The schematic on that link is similar to my version A

the problem with that version is that if the MV pot fails and the wiper disconnects

there isn't a leak resistor connected to the grid of the power tube

Franco

[Willabe]

Sluckey why don't you like K's example B wired with the pot backwards like in a tweed Deluxe?

[sluckey]

p.s.: Am I wrong if I think that the 2.2M resistor is useful only if there is a fixed bias but not so much if there is cathode bias ?

In your circuit you have a 2.2M parallel to a 1M which is really equal to a 687K.

The resistor between the pot wiper and pot bottom lugs serves the same purpose for the fixed bias circuit and the cathode bias circuit. IT PREVENTS LOSS OF THE GRID RETURN PATH. When you lose the grid return path the grid becomes an open circuit and bias cannot be established between grid and cathode, regardless of the type bias that is used. When bias cannot be established, the state of the tube is unknown, but usually cathode current will run wild and destroy the tube if it's a power tube. The same happens with preamp tubes but there is usually a high value plate load resistor that will prevent tube destruction.

The 2.2M value that you often see in the fixed bias PPIMV circuit is chosen because the 2.2M does not change the value of the pot very much. That's fine because all you want to do is provide a path for the bias voltage to still get to the grid if the pot fails. Since there's usually no grid current flowing, the full bias voltage still gets to the grid to prevent meltdown.

I chose a lower value resistor for the cathode biased amp just to keep from greatly exceeding the value of the max recommended grid return resistor. If you used a 2.2M in a cathode biased circuit and the pot failed, you would have a 2.2M grid return resistor. That exceeds the maximum recommended value. I doubt any harm would come to the tube, but I don't know for sure. I feel a lot safer using a much lower value resistor.

Phew! Having said all that, I don't have a PPIMV in any amp I've ever built or owned. 

[sluckey]

Sluckey why don't you like K's example B wired with the pot backwards like in a tweed Deluxe?

Because it's wired backwards. 

It's just wrong. That pot changes volume by changing the load on the previous tube's plate. I don't like that idea. It's illogical.

My 5E3 has the pots wired backwards. It's still wrong. Look at how the two pots interact. Leo was just saving two mixing resistors by wiring the pots backwards. He didn't do it because of some magic mojo. Who'da ever thought people would like that crazy interaction!

[kagliostro]

In your circuit you have a 2.2M parallel to a 1M which is really equal to a 687K.

Errr .... no, in my circuit I've a 2.2M (safe resistor) || 1M (MV pot) || 1M (grid leak resistor) equal to ~407K (Version D)

So, if I've the 1M grid leak resistor and the amp is a cathode bias, is important to have also the 2.2M ?

Grazie

Franco


EDIT: OK I respond by myself, if there is the 2.2M resistor and the wiper fails I've 687.5K as leak resistor, if there isn't the 2.2M safe resistor I'll have 1M that is a higher value that is way more far from the max admitted value ..... 

[kagliostro]

But now I would like (for knowledge) to know why version C isn't a good way to do

Grazie

Franco

[sluckey]

Version C wastes a coupling cap. Leo would fire you for that! 

[Willabe]

Sluckey why don't you like K's example B wired with the pot backwards like in a tweed Deluxe?

Because it's wired backwards. 

It's just wrong. That pot changes volume by changing the load on the previous tube's plate. I don't like that idea. It's illogical.

My 5E3 has the pots wired backwards. It's still wrong. Look at how the two pots interact. Leo was just saving two mixing resistors by wiring the pots backwards. He didn't do it because of some magic mojo. Who'da ever thought people would like that crazy interaction!

Ok then. 

[kagliostro]

No, seriously, version C arrangement is detrimental to the good performance of the circuit ? 


Obiously I'm not Leo and building only one amp I can afford the cost of one more cap   


Franco

[PRR]

I'll take version B, with the "backward volume pot".

Its sin is that it will distort as it is turned down. But is that a bad sin for your purpose?

All of these only work if you can reduce level enough to clip in the driver before total gain gets too low to play well. That's maybe why MV SE is rare: typical SE stages are so easy to drive you have to choke the life out of the signal-chain before the output power drops.

I'll also note that you could TRY several variants with a hot iron in a half-hour. Bench racing is fun, but the real test is on the track.

[sluckey]

I told you why I don't like C. Others do. It's an opinion type question. C is not detrimental to good performance.

[PRR]

> it's wired backwards.  It's just wrong.

I agree... but there's two sides to every rock.

It looks "backward" because you think low-Z source hi-Z load.

If you "backward" your thinking, hi-Z source lo-Z load, the backward pot makes perfect sense.

It was actually quite common in early transistor designs. Transistor gain focused on current. An unsophisticated transistor chain has ~~10K sources (collector resistors) and ~~1K loads (base inputs and bias networks). Signal levels were quoted as uAmps.

Of course these signal currents were computed, rarely measured, because current means breaking the circuit (before good clamp-on current meters). I think designer and debugger sanity (and falling prices) led to the present idiom of low-Z sources, hi-Z loads, and "forward pots".

There are more side effects (backward AND forward) when Zs are not truly high or low. In the Original Application, a Potentiometer was fed from a battery (low Z) and wiper sensed with a balanced bridge (infinite Z). The electrical ratio is exactly the mechanical ratio (assuming uniform resistance of track). But we know that a non-infinite load "curves" the "taper". The reverse pot also curves when it does not see zero and infinity.

In the DeLuxe, the 1st stage plates are moderately high Z. The mix-point is far from infinity being two pots parallel. I believe this "backward" connection gives as low loss as possible, gives semi-log (really Square) taper from linear pots, and saves two resistors. Downfalls are several. Bass falls-off at low settings. At high settings interaction is rampant. Brite cap on one side adds more odd effects. Full-up it becomes a Plate Mixer which just may be going nonlinear at high imput levels. Two resistors would give a much more predictable (non-interacting) mix system, and in this case with similar gain. But that's 25 cents, several eyelets, and the wacko network was not hurting sales. It even seems to be beloved.

[jjasilli]

Ditto to sluckey Reply#3 and labb Reply#5.  What's good about Version C is the fixed grid leak resistor for protection of the power tube in case of failure of the MV pot.  Both those replies account for that (as does Version D).  (... with the AX84 solution being particularly clever.) However, Version C could have a simpler circuit overall.  The fixed grid leak resistor, in parallel with the nominal value of the pot, should equal a value the power tube would like to see in terms of signal.  (As already pointed out, the the fixed value of the grid leak resistor should satisfy the operational need of the tube.)

[tubenit]

This is sort of a half a PPIMV idea?

With respect, Tubenit

[kagliostro]

Ringrazio tutti per i loro interventi

As PRR say, this matter will worth some practical investigation

Franco

[shooter]

this matter will worth some practical investigation

K, I'd start with a drive though the countryside, a stop in a local café, espresso, a croissant, more driving, a nice meal in the next town, go home and well, maybe 10minutes on why no body really needs
 an SE MV

[kagliostro]

Why do things as all other do ? 

All of these only work if you can reduce level enough to clip in the driver before total gain gets too low to play well. That's maybe why MV SE is rare

so the preferred version for tests seems version B    (despite the fact that my preferred - on paper - is version C)

Franco

[sluckey]

so the preferred version for tests seems version B

That depends on who you ask. 

[kagliostro]

Steve I have not forgotten your opinion

and I have your opinions in large consideration

What PRR told encouraged me to give it a try, I've understand the reason for which you don't like that option

see my reply #2

p.s.: actually solution B has the problem that the load of the preceding gain stage isn't constant and this I can suppose can give problems

So

Version A (despite it is bad engineered) is largely used (give a search with google and you'll see)

Version B has the problem that it vary the load of the preceding stage

Version C is rarely seen and waste a cap, I've seen only two or three schematics (one is the Weber AB200 Kit) that uses it

Version D is a more weighted solution and seems a good one (also if some, may be far, risk is still present)

Each of this solution seems to have a reason (may be version A reason is only that it is a very simple plan)

when PRR pointed to the fact that the reason for a MV is to try to reach distorsion of the preceding stages at low power

(the reason for which I've think to put on one)

I decided, why don't try to keep profit from the intrinsic defects of version B ?

I'm still (secretely) loving version C, but as PRR say it will be not a long time to have some test of the different possibilities

"That's All Folks!" 

Franco

 

[sluckey]

Please give us your opinion when you have completed the comparison tests.

[kagliostro]

This will happen not in a small time

(because of some problems related with the material construction of the amp that didn't depend on me)

but I'll report for sure

Franco

[kagliostro]

Well, due to health problems of my father had suspended work on the amplifier with 1625, now he feel better and I'm still here

we stopped on this version of MV for SE amp



the 2.2M resistor was added for safety against wiper failure and the 1M resistor as grid leak resistor

looking to the scheme (version D) I was thinking to use only one resistor of adequate value (feasible ?)



One other option came to my mind (may be an odd option, because it uses a dual ganged 1M pot connected in parallel, to avoid problems if one wiper fails)



and, this solution that seems to me a good one (you can see something like this on the Egnater Rebel 20)



Please, what do you think about this 3 different versions ?

Is correct version E and, expecially, what about version G ?

Many Thanks

Franco

[PRR]

> now he feel better

We are very glad to hear that.

FWIW, my dad is scheduled to get a thing done to his heart. Trivial, but they keep him overnight in case they screwed-up, and that is also my parents' 65th wedding anniversary.
___________________

Version G shorts-out the triode amplifier to get zero volume.

At low volume the triode will distort badly.

Yes, if you want a power turn-down so it distorts at low power, this may be good.

Redundant resistors, or added pot-elements, seems excessive.

The rating for 807 (base type for 1625) is 470K in self-bias.

6SL7 with 100K is about a 50K source.

Split difference 50K to 500K is 150K.

I would think a 250K or 500K pot, and 500K fixed resistor, would be fine.

[kagliostro]

Many Thanks PRR

I'm really glad for my father

My best wishes for the intervention of your father, I hope that everything will be resolved quickly and for the better

---

This I've understand well

Version G shorts-out the triode amplifier to get zero volume.

At low volume the triode will distort badly.

Yes, if you want a power turn-down so it distorts at low power, this may be good.

Redundant resistors, or added pot-elements, seems excessive.

The rating for 807 (base type for 1625) is 470K in self-bias.

---

This I think I've understand

6SL7 with 100K is about a 50K source.

but I think I'll adopt the value suggested by Steve (used in his RockOla) rp = 270K -- rk = 4.7K (and I'll add the Fender 47R for NFB - all is based on AA764 VibroChamp)

excuse me if I didn't report the correct value on schematic

With rp = 270K  will be  ~135K the "source" instead of 50K ???

---

Of this part I've understand only your council for adequate values, but not your math

Split difference 50K to 500K is 150K.

I would think a 250K or 500K pot, and 500K fixed resistor, would be fine.

---

As 1625 would accept a grid leak of max 470K, I've think to use a 1M MasterVolume pot and a 680K grid leak resistor

(1M || 680K = ~405K and this is the max value, when the wiper of the pot is connectec to pin # 3 of the pot)



---

About distortion, yes, I would like a bit, but not very much (so not sure about the odd connection of the MV), I'm planning to use an OD on the pre, the MV is there also to have more control on the power

Grazie

Franco

[HotBluePlates]

Your version H looks good.

My argument against version C is not that an extra coupling cap is used, but that the bass response changes from a low-master-volume setting to a high-master-volume setting.  When the master is at max, the two coupling caps are in series and the effective capacitance is half that of a single cap.  At low master settings, the pot's resistance isolates the two caps.

So you get more bass response at low volumes and less at high volumes.  This could be a good feature, but you'd likely need to develop the cap values experimentally.

[kagliostro]

Many Thanks HotBluePlates

When the master is at max, the two coupling caps are in series and the effective capacitance is half that of a single cap.

Actually, I know that with a PPIMV in fixed bias amp, where a cap is to be used between the pot and the power tube, the value of the cap preceding the PPIMV is to be doubled and the cap following the PPIMV is to be of the same value as to avoid the change to 1/2 of capacitance when the two capacitor are in series

Also, in an improved version ov the PPIMV, that I've seen, the capacitor between PPIMV and the power tube is indicated to be very large (1uF) and in the Collins PPIMV there is a resistor (value to be establish experimentally) placed in series between the wiper and the cap that feeds the power tube with the purpose to minimize the frequency response changes

But if you don't point to the frequency response I didn't remember to pay attention to this thing





Is strange that many people who add a PPIMV didn't care about this details and build it using capacitors that are simply equal to the existing coupling capacitors

(an interesting reading on this thread) http://el34world.com/Forum/index.php?topic=16511.msg162515#msg162515

Ancora Grazie

Franco

[HotBluePlates]

Well, the idea which came to mind was blocking distortion due to grid current once the driving signal peak exceeds the tube's bias.

The grid current tends to charge the coupling cap and shift the bias, cutting off the tube after the big peak is over.  Large capacitance (C) and/or large grid leak resistance (R) means this overload recovery takes longer.

The usual cure is to reduce C (but that cuts bass response), or reduce R (but this lowers the gain of the previous stage due to a smaller a.c. load), or both.  But since a lot of bass in a distorted sound is usually not pleasant, maybe reducing bass is not so bad.

It would be best to breadboard and see if you have any issues with the bass response or blocking distortion at high volumes.  If yes, you tweak the related parts to land at a compromise which sounds best.  It's hard to predict "best sound" for an original design; by ear is usually best.

[kagliostro]

Well, all is very interesting ... but

HotBluePlates, your post was intended for this thread

or for this other http://el34world.com/Forum/index.php?topic=21206.0 ?

Ciao

Franco

[TIMBO]

Did we find a solution......

[tubeswell]

Did we find a solution......

I know it an old thread, but finding is the fun part, so... here's another kick-around concept (just because I'm bored)


Looking for a way to incorporate MV without losing drive to the output tube, whilst preserving the '2-stage' pre-amp concept - I thought to chuck a DC-coupled CF buffer in there borrowing heavily from the SVT driver, with ability to manipulate the CF operating point through adjustable bias. The CF setup and indicated voltages are my best guesstimate  - it might work, but should be easy enough to tweak to get the right (0V) target g1 idle voltage for the output tube. It could lend itself to all sorts of output tubes  or several parallel output tubes. Totally untested and quite likely hair-brained.

[Williamblake]

I guess a grid leak resistor with current will not give you zero volt.

[tubeswell]

I guess a grid leak resistor with current will not give you zero volt.

The output tube grid leak is through the CF load resistor and then through the 12AU7 bias adjustment circuit, and also through the -VC source to ground (which is only shown as -VDC on that schematic). The -VDC source would be set up to give ballpark voltages needed to get the 12AU7 cathode idling at 0VDC, and you'd use the 12AU7 bias adjustment to fine-tune in order to get 0 VDC on the cathode/grid, depending on the specific vagaries of the 12AU7 triode you happen to plug in.


(My ballpark guesstimate for the -VDC source is -100V, based on a +VDC source of 300, and figuring that you'd bias the 12AU7 with just enough cathode current* to get the cathode idling at 0VDC, without exceeding the 12AU7 plate dissipation. Having the 12AU7 cathode idling at 0VDC, it wouldn't need any/much heater elevation)


* the 12AU7 cathode current would include a component which is output tube grid leak current

[TIMBO]

Thanks guys, This is my schematic with an added 1M leak resistor.
Works fine without the resistor, haven't tried it with the resistor.

[kagliostro]

Ciao Tubeswell

Very interesting idea

--

if the add of a tube is acceptable to achieve MV

will this kind of tube be interesting for that application ?

https://dalmura.com.au/static/6BE6audi.pdf

(Signal on grid 3 and control voltage on grid 1)

Franco

[tubeswell]

Ciao Tubeswell

Very interesting idea

--

if the add of a tube is acceptable to achieve MV

will this kind of tube be interesting for that application ?

https://dalmura.com.au/static/6BE6audi.pdf

(Signal on grid 3 and control voltage on grid 1)

Franco

Well, keeping in the spirit of pointless discussion about adding MV to a SE amp for guitar, that tube looks interesting. Have you got one?

[kagliostro]

Yes, I have some


Not for sure they are the original 6BE6 tube, but surely I've some russian version (unfortunately at the moment I don't remember well nor if I've some original nor the russian designation for the tube)


Franco

[tubeswell]

Looks like it could work with variable DC source