Check my schematic help mod / tweak please SE 6v6

[HiFiFreq]

Just built this amp and would like some input please . It is a single ended 6v6gt output
with 12ax7 pre and SS rectification . As it sits the tone is a little ratty sounding , harsh .
Kind of a princeton amp but with different cap and resistor values . I used what was in arms reach and junk box parts .
Included is a schematic with voltages and values . Thanks in advance TDPRI !

 

[jsnwhite619]

Is the 360R cathode resistor biased too hot? I usually go with a 500R for my 6v6 Champ/Princeton stuff, though 470R was stock I think.

 

[HiFiFreq]

It's a junk box Princeton , maybe . I punched the numbers into RobRobs bias calculator and it seems to be about 93% of PD . If I am using the calculator correctly . Also the plate voltage is lower than the screen voltage . May be because of my power rail . Do I need to drop some more voltage between the first and second node ?

 

[chas.wahl]

A 5F2-A has a 10k dropping resistor, you've got 1.5k (15%); if I were you, I'd raise that so that the voltage at the 6V6 screen grid is appreciably lower than that at the plate, so you don't burn it up.

The caps on your tone circuit are 2/5 of what they'd be on a 5F2-A. There are a bunch of other discrepancies in both coupling cap and resistors, on the order of 2x or 1/2x difference. Obviously, if you use other values, then the amp will sound different from a 5F2-A, maybe some of that difference is what's "ratty" and "harsh". Another approach might be to stick to the 5F2-A values to begin with, and incrementally change things, seeing what happens with each one. Or, maybe you just don't like the noises a single-ended amp makes.

 

[andrewRneumann]

My impression:

1. You need some kind of grid leak on V1A. Maybe you didn’t draw it on the schematic. The typical 1M to ground is to be expected.

2. Harshness
It’s a subjective term. I will tackle it from a distortion point of view. With two fully bypassed cascaded 12AX7 triodes and no negative feedback, you are going to get some serious overdrive. Consider temporarily lifting the cathode bypass cap on the 2nd gain stage and see if that is to your liking. If that’s not enough, try out NFB running back to the cathode of V1B with no bypass cap.

What OT are you using? That one has quite a lot of DC resistance and that is why your plate voltage is lower than your screen. Possibly it is too small for the job. You are dissipating about 1W just in resistance of the OT itself. If your OT is saturated with DC current, that can cause distortion too. Some people like that sound… If you have a slightly more capable SE OT on hand, it might be worth a shot.

 

[printer2]

120k grid stopper on the input, just adds noise, 33k more or less as Fender did, as mentioned above, you need a grid leak resistor. Throw a grid stopper on the 6V6 also. I would drop some more voltage for the preamp, less hum. The 6V6 is a little hot biased, maybe a 470R? A little less current might make the OT a little more happy. A screen resistor also.

 

[FenderLover]

Not a real fan of the 220K plate resistors either. Sometimes the junk drawer can get you going, but those passive parts are pretty inexpensive. Spend $10-20 on common values and you'll be able to start closer to a 'known good' state. Tune from there.

 

[cobaltu]

I'll echo andrewRneumann. I think that the there is just too much gain in this amp. Just a side note - distortion tends to sound best when it is produced subtly over many stages (like in a high gain amp). It it happen relatively all at once (like in an amp with only 2 stages capable of overdrive) it is usually 'ratty' sounding.

What I would do:

1) Definitely take the 6v6 cathode resistor down to ~500ohms, and go from there

2) Lift the cathode bypass cap on the second gain stage (this is also necessary for another one of my recommendations)

3) Increase the first cathode bypass cap to ~4.7uF-22uF, otherwise the amp will sound quite thin with just that 1.3uF you have currently.

4) I would also add some negative feedback, with a 2k cathode resistor on the second stage, 22k-33k resistors would be a good starting place. (Remember, in order to apply NFB you would have to feed it into the cathode of the second gain stage. But if you bypass that stage with a cap it would work, it juts shunts any feedback to ground).
I have an Allen "Classic 10" (another 5f2a variant) and I must have loosened the negative feedback on that amp 4 or 5 times and I always came back to the stock 22k value. For me it was always the same problem - the amp sounded to brash and ratty with less NFB.

There are some other things you could do too, but these I think are some good starting places. Good luck.

 

[cobaltu]

Oh, I also agree with printer2 as well. The 120K at the input is just adding noise and possibly truncating treble too early. Again 33k would be a good stating place.

Another side note about shaping distortion in an amp - it's typically better to cut bass early and cut treble later in a circuit, especially when overdrive is a goal.
Cutting the bass earlier keeps the circuit tight and the bass frequencies from building-up too much and making everything sound like a wet fart.
Cutting the treble later allows the upper harmonics to build-up and play a role from the compression and distortion as it heads through the circuit. But by cutting it as late as conveniently possible it cuts out the harsher stuff that isn't necessary for a guitar amp.

 

[HiFiFreq]

Thanks for the suggestions . I will start with the power supply dropping resistor first and then the K resistor on the 6v6 will probably be the second change . I wont be able to start these changes until the weekend . I scored a variac at the thrift store yesterday and have been working on that today . Looks like someone dropped it so I've been straightening out the sheet metal of the enclosure . It works but it needs a high dollar switch . I'll end up paying twice as much for the switch ($50.00) as I did for the variac.

 

[HiFiFreq]

Some progress . I should have described the tone as ragged maybe . I changed the first dropping resistor to a 5k resistor . That helped clear up the raggedness immensely and brought my screen voltage down under the plate voltage by about 10 volts . I removed the bypass cap from the second pre-amp stage and installed 25k NFB . That quieted down some noise . I added a 4.7uF bypass cap to the first pre-amp stage . I also used a 500ohm resistor instead of the 360ohm resistor on the 6v6 tube. The amp is sounding much better now . I did not mark it on the schematic but there is a 1meg grid leak resistor on V1a input . I will take some voltage readings to plug into the RobRob bias calculator now .

 

[HiFiFreq]

Using Rob Robinette's bias calculator I come up with 10.3 watts PD . 73.6% of max 14 watts .
6v6gt plate to cathode voltage is 284vdc . Voltage at pin 8 is 19.2vdc . 500 ohm cathode resistor .
Where could/should an amplifier like this be biased ? I thought I needed to be closer to 100% .

 

[2L man]

Grid 1 leak resistor 560k is more than double the original schematic and exceeds 6V6 max 500k! You should test about 250k.

Original schema does not have Screen resistors but using for example 470 ohm "tame" power tube slightly.

Also g1 grid stopper is often installed...

 

[2L man]

Using Rob Robinette's bias calculator I come up with 10.3 watts PD . 73.6% of max 14 watts .
6v6gt plate to cathode voltage is 284vdc . Voltage at pin 8 is 19.2vdc . 500 ohm cathode resistor .
Where could/should an amplifier like this be biased ? I thought I needed to be closer to 100% .

Where did you set the OT impedance? Plate voltage and OT impedance = loadline/operating point can be set many ways. On typical champ the OP is to the right and output clips already quite low power and produce lots of 2nd harmonics.

For 7k OT impedance loadline comes about this. I did use 3.4mA less plate current because Screen voltage is high and no screen resistor which keep Screen voltage higher when its direct B+2 is filtered.

For SE amps I set the Headroom so that 2nd and 3rd harmonics are about the same. Although they are just estimates I get one variable which make it easier to compare loadlines.