How to determine filter caps values?

[Patrik]

Greetings,
I am a novice here and am currently in a process of building my custom amplifier for the first time. I made a few replicas, Fender 5F2a and Vox AC4, but wasn't fully satisfied with the sound and feel, and am now designing my custom one from scratch. The goal is to have an amplifier that is as simple as possible, single ended and only gain, tone and volume fir controls, light and portable, made so it can still do effects like echo and fuzz, and rich in midrange and lower highrange, being able to saturate it only using the volume knob on a guitar.
Now I settled on EF86, first stage, ECC83, second stage, EL34/6L6/KT66, power tube, and EZ81 rectifier. I had no problems designing the actual amplification circuit but power supply is another matter. I don't understand how filter caps values ate calculated. I don't want to use choke, to save weight, and will be using Hammond 372BX or 373X as a power transformer and Hammond 125DSE as an output transformer.
Any help on that matter will be greatly appreciated.
Thanks in advance,
Patrik 

[Latole]

https://el34world.com/Forum/index.php?topic=15436.0

http://www.valvewizard.co.uk/smoothing.html

http://www.bustedgear.com/faq_Amp_recap.html

[Patrik]

Thank you.

[Latole]

Thank you.

You are welcome.

[ac427v]

Here's some controversial thoughts:
I have done some thinking about the single ended amps I built and why I was never fully satisfied with the sound and feel. I think I always missed the solid bass frequencies that come with almost any push-pull amp. I tried bigger transformers, 6L6GC power tube, higher supply voltage, bigger speaker and cabinet. Heard great mids and distortion. But empty bass. Then there is the weight problem. Single ended transformers are significantly heavier than push-pull of the same wattage. Many builders here seem happy with single ended designs. I predict they will share their experiences too. Their dream amp sound is different than mine. Based on your brief post I'm wondering if you might be happier with an amp with push-pull power tubes.

If I've misread your interests and you want to go forward with your plan, consider using a 5xxx rectifier.  That would better match the heater options of your intended PT. Happy building!

[Latole]

I build a single end amp 15 years ago because I have lots of "very good" PP amps. I don't need another one.
And it is a fun project,

[Patrik]

I have played every type of amplifier there is but wasn't fully satisfied with anyone due to my tendency to control gain and volume via tone on volume on guitar, and many modern amps are built so they can take all sorts of pedals without a problem. The closest I got to the amp I like was a Dynacord KV10, I liked the response a lot, but I didn't like the overdriven sound even when it was retrofitted with a Celestion G10 Greenback. For now I just want to experiment a little, just for the fun of it and if having fun isn't enough of a reason I don't know what is.
Also if I will be at least semi-satisfied I will still try with the same preamp, that I moddeled after VOX AC15, combined by a tad of 5F2a.

[pdf64]

http://www.duncanamps.com/psud2/

[PRR]

...I don't understand how filter caps values ate calculated....

Mostly not calculated; use the same values which worked before. Adjusted if they didn't work so well (or "too well").

Light weight and SE 6L6/EL34 may not go together. (Commercially this was always a very rare combination, even in fixed-install amps.)

[brewdude]

I have a SE amp I built... actually, I have rebuilt it many times, and I just don't like it as much as the push-pull amps I have built.
I used a 125ESE OT.  I've used a 12AX7, 6SJ7 (pentode) and 6L6 in the latest incarnation.  I've changed things many times and don't remember where the circuit stands, but I don't desire to play it much.  I hope you have better luck.

[shooter]

I've used a 12AX7, 6SJ7 (pentode)

that's enough gain to trash PSE KT88's
try setting the VAC rms to about 1.5X the cathode vdc


is the OT UL? (I don't want to look)

[brewdude]

I don’t want to highjack this thread...
It was an experiment.  My original goal was to get the pre-amp pentode to overdrive and then amplify the distorted signal with the class A PA.  There were some good things and some bad things that I learned.


Can you explain setting the VAC?


Once again, sorry if this is distracting from the OP’s thread. 

[Colas LeGrippa]

If you want to use a rectifier tube, you MUST follow the rectifier tube specs concerning the filter caps values.
Check out the data sheet for that particular tube

[Patrik]

Thank you guys, You have been much help to me and this is is what I came up with.

[Latole]

Nice sketch,

Why EL34 ? Is it not too powerful and expensive tube for low power amp ?

[Patrik]

I would like somewhere around 10W of output and tubes like EL34 and 6L6 in SE configuration can supply that. And they are rather cheap here in Europe.

[Latole]

I would like somewhere around 10W of output and tubes like EL34 and 6L6 in SE configuration can supply that. And they are rather cheap here in Europe.

How can you know you will have 10 watts ?
EL 34 is cheap here too.
EL84 is cheaper but only 6 watts

[Patrik]

They are rated to max out at 10W.

[Latole]

They are rated to max out at 10W.

If you feed it with the right preamp.

With only one preamp tube you will not have more than few watt

What you say is like, I use 100 watts speaker so my amp power will have 100 watts  output

[Patrik]

I don't think that number of preamp tubes has such an effect on the output power of the amp, since most amps have only one preamp in SE and one preamp and phase inverter in PP configuration. I am far from an expert when it comes to how the amp works, but I still think that the preamp I designed will be sufficient to power one EL34 up to 10W. 

[Latole]

 since most amps have only one preamp in SE and one preamp and phase inverter in PP configuration.

I never see these amp with more than 2-3 watts output.
Can you show me one ?

Update ;

This one ; 10 watts :   1 X 12AX7 and 1 X 6L6  or EL34

I have always somthing to learn. 

http://www.allenamps.com/classic10.html

[pdf64]

Thank you guys, You have been much help to me and this is is what I came up with.

Your coupling and decoupling cap values are ginormous, orders of magnitude greater than is desirable for an instrument amplifier, more appropriate for a seismometer 

There are many instrument amps with a single stage preamp (in one channel at least), just requires reasonable gain from the power amp and a sensible definition of which stages fall under which category.

[Latole]

Power supply voltage must be ( very ?) high too.
Not sure 300 volts is enough and enough curent with this rectifier tube

[Patrik]

I can go up to 350 with EZ81 what should be sufficient, but EL34 is rated for 250V as I can see on the specsheet, so I don't see a need for a higher voltage transformer, please correct me if I am wrong. About coupling caps, I am using the same value as is standard on Fender amps, but will be happy to change them if you thing that change would help.

[Patrik]

And I got the idea for for such an amp from Weber Maggie.
https://tedweber.com/maggie-c-kt/

[Latole]

And I got the idea for for such an amp from Weber Maggie.
https://tedweber.com/maggie-c-kt/

[Patrik]

Yes.

[scstill]

Thank you guys, You have been much help to me and this is is what I came up with.

Why do you have two different plate voltages on V2?
Typically same tube would use the same plate voltage.
Also why did you design with three gain stages before your power tube?

[Patrik]

For the plate voltage I don't really have an answer. Probably just becouse I am used to seeing resistors between filter caps. And I would like a bit more gain than a single 12AX7 can provide.
I am not an expert in designing amplifiers what is the reason why I begun this post and would appreshiate any suggestion and opinion.

[pdf64]

Thank you guys, You have been much help to me and this is is what I came up with.

Another thing is that it's essential to provide a dc path between a tube's control grid and cathode, ie the grid must be 'referenced' to the cathode. Usually that's done via a gain pot or grid leak resistor, with cathode bias, both would be referenced to circuit common 0V.
Your first and second stage's grids are missing such a reference.

The 100k input grid stopper will cause a lot of hiss; pentode input stages generally don't need a grid stopper.

Regarding the coupling cap values, they work with the resistance in the circuit to provide a high pass filter, whose corner freq is determined by the cap's value C and (total effective) resistance R values. Such that F = 1 / (2 x pi x C x R)
In the old days, a useful 'rule of thumb' (ad hoc guideline) was that for a 'full bandwidth music' (50Hz-15kHz) reproduction, C (in uF) x R (in k ohms) should = about 10 (to my recollection, the 20Hz-20kHz 'hifi' standard was more of a 70s transistor era thing).
That gives the filter's -3dB corner freq of about 16Hz, so minimal attenuation of any guitar bottom end (a bass amp may need a bit more, eg CxR = 20).
These days, using an online calc is easy, eg http://www.sengpielaudio.com/calculator-RCpad.htm but the old CxR=10 is still a pretty handy 'hack'.
So if a stage is feeding a 1M pot (ie 1000k), a 0.01uF coupling cap may be a good place to start. If, having tried that, the bottom end needs 'tightening up', the cap's value might be reduced to 0.001uF; maybe that's too far and 0.0022uF will be settled on, or there's still too much bass, and perhaps 470pF will work out best.
With guitar amps, they tend to be overdriven, and if the above 'CxR' value is too high, a big problem is that if the stage being fed gets pushed into grid conduction (which typically defines the onset of overdrive), the Vac of the signal becomes rectified and the grid's reference Vdc gets pulled negative below 0V. This is referred to as 'bias shift'; some degree of it is inevitable and sounds fine, but too much and bias gets closer and closer to cut off, the tone gets farty, until the whole signal gets momentarily choked out, referred to as 'blocking distortion'.
Old Fender designs often used a CxR of about 20, and perhaps up to 50, hence eg a Tweed Deluxe can get farty and typical mods include reducing the coupling cap values.
And over the 60s, as Marshall gradually tweaked the 5F6a based design they started out with, some key coupling cap values dropped in value, eg the CxR of the coupling caps to power tubes reduced from 20 to 5 (JTM45 to JMP 1987 / 1959).

Aiken's tech info pages are really useful reference source, written by a proper EE but reasonably easy reading / not too daunting, eg http://www.aikenamps.com/index.php/designing-common-cathode-triode-amplifiers

[Patrik]

Thank you a lot. I didn't even notice problems with coupling. Thanks a lot.

[Patrik]

I hope that fixes problems with catode coupling. I also changed the coupling caps.

[sluckey]

I hope that fixes problems with catode coupling. I also changed the coupling caps.

Still missing a grid return resistor for that triode.

And 1000µF seems excessive for a cathode bypass on that EF86. Where did you get that value?

[Patrik]

I got that cap value from diagram linked below (#1). The second one is what I think might be better based on your comment.

[pdf64]

Thank you guys, You have been much help to me and this is is what I came up with.

Regarding the HT supply -
From p4 of https://tubedata.altanatubes.com.br/sheets/010/e/EZ81.pdf when feeding 300-0-300 into an EZ81, the Vdc at the reservoir cap is more likely to be about 350V.
With the first node / reservoir as 'E' and the last node as 'A', by the normal conventions on such matters, your HT nodes are labelled backwards 
The last node (A) is missing its decoupling cap.
The EZ81 can cope with a higher reservoir cap value; why not make use of that capability, as with SE amps, hum from HT ripple is a problem. Hence I suggest 47uF for the reservoir cap, or at least 33uF.
For other values, I wouldn't use less than 22uF unless specifically copying a vintage design.
Note that SE amps greatly benefit from the use of a CLC / pi filter in the HT supply to the OT / plate, eg 5E1 https://el34world.com/charts/Schematics/files/Fender/Fender_champ_5e1.pdf as the reduction in HT ripple, and hence hum contamination of the output signal, can be significant.

[pdf64]

I got that cap value from diagram linked below (#1). The second one is what I think might be better based on your comment.

Those circuits are intended hifi type applications.
This calc for cathode bypass is accurate and complies to the method in RDH4 http://bmamps.com/CapCal.html

[Patrik]

It might be easier to put another 12AX7 in the amp that would replace the EF86. Then I could also move controls, so that gain would be between first and second stage and tone between third and fourth. How does that sound to you? My reason for using an EF86 is purely due to it being a bit more exotic than any old 12AX7. But I am much more familiar with with double-triode tubes than pentodes for preamp. Maybe it would be good idea to use 12AX7 so that I don't try too many new things at once. I think that making a self-designed amplifier is enough of a job in itself.

[Patrik]

So, there is a completly revised diagram of the amp. I replaced EF86 in the first stage with 12AX7, put gain and tone control between first and second stage, but power amp remained unchanged. I also made aa few changes to power supply and heater wiring.

[sluckey]

Still missing a grid return resistor for that triode.

actually missing two now.

[Patrik]

Where? Do you mean grid leak resistor?

[sluckey]

triodes 3 and 4.

[Patrik]

Honestly I have no idea where to put them. Help please.

[shooter]

'tween the coupling cap and the grid

[Patrik]

Thanks, would 68k do? Also, does each triode has to have its own grid stopper resistor? I thought that one on the input would be enough.
Update: just found out that 40k would do.
http://www.valvewizard.co.uk/gridstopper.html

Thanks for noticing.

[sluckey]

Not talking about grid stoppers! Connect one end of a 1 meg resistor to the grid pin 2. Connect the other end of that resistor to ground. Call it a grid return resistor. Some people call it a grid leak resistor. Repeat for the other grid pin 7.

[Patrik]

Ok, thank you.

[AmberB]

I've built 2 different SE output amps, both using a 6L6 for the output tube.  Both amps use a PT and OPT from small PA amps.  Both amps use a 5Y3 rectifier, as per the original amp the transformers came from.  Both amps are a variation of the Fender Champ circuit.  On the first amp I used a bass and treble control instead of a single tone control.  On the second amp, I used a treble, mid, bass control setup.
I actually like the sound of the first amp better through the same speaker.  Both amps were built as little heads, not combo amps.

I can see where a simple tone control could have an appeal for a simple amp.

[Patrik]

The main reason for gain, tone, volume control setup is due to my playing style, that I control my volume and drive via controls on the guitar, without any external effects, sometimes I use tape delay or fuzz but it is only for tone and not drive. The benefit of a simple contol setup as I see it is also for gigs. Before all that corona I used to play live almost every weekend using either Dynacord KV10 or Echolette Echobell. KV10 has only volume and tone and Echolette has volume only with push-pull treble cut on each channel, has four. The other guitarist in a band had some Fender, I think it was Frontman 212, and used at least 30min setting up his pedalboard and making tweaks to his sound, I was done in 2min. So for me simpler is better.

[AmberB]

That does make sense to me.

[DummyLoad]

I would like somewhere around 10W of output and tubes like EL34 and 6L6 in SE configuration can supply that. And they are rather cheap here in Europe.

How can you know you will have 10 watts ?
EL 34 is cheap here too.
EL84 is cheaper but only 6 watts

see attached. pg. D1, closer to 11W with the published spec.   

--pete