questions on input stage

paba · 2010-09-04 8:45 pm

Hi all,

I am think about making a change to the input stage of my mono bloc amps.
Changing from cathode bias to a battery on the grid as shown in the two PDF below of before and after.

My question is what would be the impact on gain and output impedance of the stage between the two options?

I have not been able to find much text book ref material on the battery grid bias. The first one of course is well documented with gain and impedance formulas including the case with no bypass cap.

Oh and feel free to give your opinion on this change (excellent, waste of time, etc..) with some back up reasoning or experience. Or corrections / additions of course.

thanks
paba

piano3 · 2010-09-04 9:20 pm

Decrease in gain/increase in ra are normally only experienced when there is feedback from the cathode, ie. when you remove the cathode bypass capacitor. I have never used battery bias(although I did once toy with the idea of putting a rechargeable battery in the cathode of the 2nd stage of a preamp) but I don't see any mechanism for feedback here. I find feedback to be the most difficult topic in audio and I feel that the day that I totally grasp it will be the day I cease to be a newbie! Hopefully one of the big guns will answer your post.

piano3 · 2010-09-04 9:21 pm

Oh!, getting rid of a cap in the cathode is certainly desirable

DF96 · 2010-09-04 10:33 pm

Assuming the bias voltage is correct, then it will have no effect on gain or output impedance.

You may have to look in very old books to learn about battery grid bias, as it is rarely used now. An alternative is LED bias in the cathode circuit, which is now quite popular. Both of these have the disadvantage that they don't compensate as the valve ages, and you need to think about what happens when the battery ages. I have not tried either.

DF96 · 2010-09-04 10:40 pm

Something I forgot to say: you get rid of an electrolytic, but add a cell, so you just swap one type of electrochemistry for another.

piano3 · 2010-09-04 10:58 pm

DF96 said:
Something I forgot to say: you get rid of an electrolytic, but add a cell, so you just swap one type of electrochemistry for another.

Yes, I hadn't thought about that! At least,though, a battery couldn't cause motorboating,could it?

DF96 · 2010-09-04 11:08 pm

No, because the battery goes down to DC - unlike a capacitor.

It is CCS bias which can sometimes cause motorboating, because when bypassed it converts a lead-lag into an LF rolloff.

rman · 2010-09-06 3:24 am

Hi.

How about using a rechargeable battery on the cathode instead of on the grid?
This works very well if the voltage is what you need. About 1.3 v. for nimh. cell.
As a bonus, the cell has no load on it, it only acts as a voltage reference and is actually kept charged by the cathode current.

I have an amp like this and it works great

Cheers.
Rolf.

rman · 2010-09-06 3:32 am

PS.

No bypass capacitor needed. gain and output impedance are the same as for a fully bypassed cathode resistor.

Rolf.

paba · 2010-09-06 6:10 am

I must admit, I expected either the gain or the impedance to change... if they are both the same then, I am really tempted to try this.

As for LED or rechargeable battery in the cathode, I considered it, but is just a different flavour of what is there now, sure probably better than the cheap polypro cap that is there now.

Anyway I will probably try this in the next couple of weeks and report back.

Thanks
paba

mashaffer · 2010-09-07 12:30 am

Is that battery bias correct? Shouldn't it be across the input signal rather than in series with it?

costis_n · 2010-09-07 12:45 pm

Sorry to hijack the thread, but I am concerned with something a little irrelevant. I am about to construct my first amplifier (integrated). Now, concerning that most grounded cathode input stages are biased at -1.2 volts or so, how is it possible to avoid overloading the input valve, when the output of my DAC is 2.3 volts peak (maximum, measured with a 1MΩ probe on the scope)? I mean, if such a signal is applied, we'll run into grid curren, even for an instant, no?

Maybe this this adjusted with the divider formed by the volume pot/grid resistor and a series resistor ? doesn't look very nice to me....

SY · 2010-09-07 12:55 pm

Not irrelevant at all, it's an important design consideration. If you're using no feedback to the cathode, you indeed could have an overload problem. Now, if you're always attenuating the input by at least 6-8dB, you'll be OK. Same if you have feedback to the cathode. But undegenerated, your overload will be something less than the standing cathode voltage.

costis_n · 2010-09-07 1:21 pm

Oh right, I forgot to mention that there is no feedback on the cathode, also that the cathode capacitor is unbypassed. And since that i plan for the whole range of the volume control to be usable, I guess I have to find something that will bias correctly in -2.5 V and 280 or so supply volts.

Thanks, SY, my thoughts, exactly.

DF96 · 2010-09-07 1:30 pm

Shouldn't it be across the input signal rather than in series with it?

No, if it was across the input it would short out the signal.

SY · 2010-09-07 1:39 pm

costis_n said:
Oh right, I forgot to mention that there is no feedback on the cathode, also that the cathode capacitor is unbypassed. And since that i plan for the whole range of the volume control to be usable, I guess I have to find something that will bias correctly in -2.5 V and 280 or so supply volts.

Thanks, SY, my thoughts, exactly.

Not so fast- if the cathode resistor is unbypassed, there's feedback to it. Is that what you meant? If so, what's the tube and plate and cathode resistor values? We can estimate overload with that info in hand.

paba · 2010-09-07 9:49 pm

So are we saying that even in my case as shown in the PDF in the first post, I'm at risk of overloading the first stage? The signal will be attenuated in the source selector/volume control unit coming before.

That's a good point then, because I saw this battery grid bias scheme on phono stage and hence the signal is either MC or MM are much smaller than in my case (tuner or CD) i.e. line level.

/paba

SY · 2010-09-07 10:18 pm

Yes, both of those stages will be very prone to overload and excessive distortion with signals greater than a volt or so. It strikes me that the paralleling is also superfluous, but maybe you had some logic behind it?

mashaffer · 2010-09-08 1:35 am

DF96 said:
No, if it was across the input it would short out the signal.

Doh, of course. Sorry for the brain fart.

paba · 2010-09-08 2:37 am

SY said:
Yes, both of those stages will be very prone to overload and excessive distortion with signals greater than a volt or so. It strikes me that the paralleling is also superfluous, but maybe you had some logic behind it?

Paralleling the two triodes in 6N1P is I assume to get enough gain or voltage swing to drive the next tube in the amp, lower the stage's Z and to get some noise benefits. The current / original input stage in cathode bias is not my design.

thanks
paba

questions on input stage

Member

Attachments

Member

Member

R.I.P.

R.I.P.

Member

R.I.P.

Member

Member

Member

Member

Member

Ex-Moderator

Member

R.I.P.

Ex-Moderator

Member

Ex-Moderator

Member

Member