idea amp

[mr.jetski]

hi all,

I'm trying to design an amp for me (nfs) I made a schem but values are missing. feel free to add what you know, change it if you must but please tell me why so I can learn. I also want to add trem and reverb. the pt will be marshall 350v-0-350v 290ma, ot will be hammond 1650a series. enzo thanks.

[black_labb]

have a look at other schematics to get ideas for values.

the way the screens come out from the pt isnt what i've seen in the past. is there a reason you did that?

[cbarrow7625]

Quick question for you. Why do you have the "ground" leg of all the potentiometers going back to the previous cathode through a cap rather than straight to ground. That is most unusual. What effect do you hope that will have on the sound?

It is hard to recommend values for something so unusual.

Also, you have a mistake in the PI. You have both of the grids for each side of the PI tied together & grounded through a cap. This is incorrect for two reasons: 1.) each grid needs to see a signal with inverse polarity from the other. They should not be tied together. 2.) Even if tying them together would work, you have crated a AC ground going to both grids. That will basically shunt all signal to ground.

I think what you intended to do was to drive the grid of the first half of the PI and make the other half a grounded grid amp (by shunting that grid to ground through a cap). However, I think you still need to provide a DC reference to that grid. You may be better off just tying the second half rid straight to ground (no cap).

[Enzo]

Black, his screens are wired "ultralinear", though we often see that without screen resistors. personally I think screen resistors are a good protective step.

I for one would like to see part numbers, so we can refer to them. If I want to call out the second tube from the left on the top row, it would be far easier to call V2. Or V1B, or whatever.

The odd pot grounding scheme won't work. Look at the upper right tube. There is no ground return path for its grid. The volume control normally would provide that path, but the cap blocks that. AS it is, that grid can now accumulate a charge and probably motorboat. The upper left tube has both - that is redundant.

Tone stacks all over!!! GO over to DUncan's site and download the tone stack calculator. That provides real time display of stack response. Experiment away there on your screen.

The common values you see in amps - for example the input 12AX7 stage has typical 1.5k cathode resistor and 100k plate load resistor - are time tested and also right out of the RCA book. Not a damn thing wrong with those numbers. SOme circuits use other numbers. You can change them, I'd just have a reason first.

SOmething I did for a guy once - he wanted me tweaking his amp and every time he wanted something different. I finally got a handfull of small trimmer pots and mounted them in place of resistors in side. "Here, you go, now YOU can twiddle the values." A 100k 0r 250k trimmer in series with a 100k plate load allows tons of variation. You could do something similar at the cathode.

The output stage is odd. You show both cathode resistor and fixed bias. If those are 1 ohm cathode resistors for current sensing, then mark them. If they are intended for biasing, then the bias supply is not needed. You could use some of each, yes, but why?

Power suplies have really low impedances. SO if you are using a bias supply, those caps to ground on either side of the bias balance pot are not needed.

cb already mentioned the PI won;t work.

PLus there is no grid path to ground there unless the lower left input stage is switched in. That volume pot covers that then.

You show two input stages selectable then a third stage that can be added in or not. I think there is a lot of potential for noises here. You should consider adding relays or more switch sections. When NOT selected, the inputs and outputs of the extra stages should be grounded. otherwise their gain will amplify whatever noises the unterminated inoputs can pick up, spreading it around inside the chassis. Might even hear radio.

Your NFB doesn;t do anything. First, the speaker output is not grounded. it doesn't need to be, but if you want some signal sampled off it, then it DOES have to be referenced to ground. Usually the bottom lead of the tranny is grounded and the three speaker taps are hot.. Your NFB usually is taken from a hot tap. If it is to be NFB and not PFB, it matters which side of the winding you use and which is grounded. get it wrong and you have built a power oscillator.

SO assuming the tap is off a hot point, then it runs to the cathode of one of the inoput stages. That means it only could have any effect at all when that one channel is selected on. When that channel is "off", then all we have is a resistor to ground off the speaker out. The way it is drawn makes me think you intended it to go to the tail of the PI.

[black_labb]

Black, his screens are wired "ultralinear", though we often see that without screen resistors. personally I think screen resistors are a good protective step.

makes sense, thanks

[mr.jetski]

idea amp 2.1

thanks all,
how is this?

[Enzo]

Still no ground refernce for the grid of the lower triode in the PI. Add a resistor.

[mr.jetski]

is this what you mean, enzo?

[mr.jetski]

idea amp 2.3

hi all,

I have put values and such in as much as my ability allows. Ill do the tone stacks later but please I really need someone to go through this with a fine tooth comb. I would really like to start to build soon. thanks

[tubeswell]

Hiya Mr Jetski

My 2CW - if you want a paraphrase inverter maybe take a look at the 5E9A schematic. (See how the 5E9A paraphrase takes the 'lower' grid from the 'bottom' of the 'opposite' output stage's grid load resistor and there is a 6k8 resistor between that and the output stage cathode ground.There is also a high voltage 100pF cap between the plates of bothe side sof the PI, which is a sort of hi-freq attentuator.)

Looks like you've got three different pre-amps switchable in and out. You might find with more than one of them switched in at a time that the signal going into the PI is way over the top.