Do FIlter Cap Values Matter in a Marshall? IN DEPTH MEASUREMENTS and TONES!!

Continuing to fine tune my own 1983 JCM800 2204 1987 50w. Today we are experimenting with power supply filtering via power tube screen supply node and filter cap. A/B test of a 50+50uF can cap for the screens and phase inverter up against a 16+16. A bit of theory and some investigations at the bench to objectively observe any difference. Lots of tones too with the 16+16uF cap in place.
headfirstampli...
/ headfirst.amplification
/ 225949122469679
/ headfirst_amplification
[e-mail]
Signal path:
Guitar - Amp - Suhr RL - Ownhammer GNR IR's - Apollo Twin - Logic Pro
0:00 Play in - 16+16uF
1:07 Intro
3:45 50uF vs. 16uA A/B test
4:32 Feel - my thoughts
5:43 Schematic - filter nodes
10:45 B+ supply line
13:08 Filter caps as low pass filters - the math
16:04 More tones
16:43 At the bench - 50uF measurements
19:58 Time lapse cap change
20:18 At the bench - 16uF measurements
21:49 Comparing the 50 vs. 16uF scope trace
25:10 Play out

[@terrydemol5354]

I'm enjoying these vids, you're putting a lot of time and effort into micro tweaking. The more pronounced overshoot or ringing with 16uF is a function of the LC low pass filter response. No ringing it is said to be critically damped. There are many different LPF responses for example Bessel, Butterworth, Chebychev etc etc. Bessell (or Gaussian) has no overshoot / ringing and for example Chebychev has a lot. Looks like there are a few things happening here simultaneously -> one is the PS 'sag' and another is the ringing from transients. I've never really thought about ringing in this part of the PS so really glad you posted this, it's given me a few ideas to try. For example, you can have the sag of the 16uF without the ringing, it's just a matter of a/ juggling the values for a Bessel response or b/ use an RC damping network. I recommend downloading Ltspice freeware and you can simulate / play with ideas, see how it responds to various IP signals before modding and listening.

[@simonkormendy849]

If you're using solid-state rectification in your amp, reducing the main supply filter cap is going to have an impact on the amp's bass response, as well as the hum levels, on the other hand increasing the main supply filter cap is going to "stiffen" the amp's response, if you're using vacuum-tube rectification in your amp you don't want to increase the main supply filter cap because it will shorten the life of the rectifier tube.

[@dr.danamplifiers1753]

As always, fantastic playing and fantastic theory. Well done mate 🤙

[@HeadfirstAmps]

Thanks Dan, really appreciate your support buddy. Means a lot.

[@Jonathan_Doe_]

You should test the sag using pulses at different frequencies, then you’ll see the depth of the sag, and recovery time, at each frequency.

[@joparkcat]

Thanks for your educational videos, the more "into the weeds" you do your tech stuff the better.

[@joshk5686]

I love that power tube swirl from that amp. It sounds really neat like a built in phaser or something.

[@HeadfirstAmps]

Hard to beat ya

[@GabeHelma]

Hold on, so there wasn't a phaser in front??? Wow i definitely thought he had a phase90 in front 😆

[@Jediroller]

Thank you for making your videos, they so cool for any amp enthusiasts trying to understand and grasp the concepts better.

[@HeadfirstAmps]

Glad you like them!

[@2204JCM]

Fun Fact: on the current fractal devices you can “virtually” adjust the filtering. Also you can select AC or DC. The AC/DC option lets you get rid of any ripple regardless of the amount of actual filtering you choose (DC). If you select AC then you can hear ghost notes if the filtering is low enough -just like in the real world.

[@HeadfirstAmps]

Amazing isn’t it, I have played around with those parameters.

[@2204JCM]

@@HeadfirstAmps Yeah, I’ve been using the Red Channel of the Bogner Ecstasy (Fractal) and its super chewy with the default settings.

[@2204JCM]

@@HeadfirstAmps Also when I turn the bias up hotter on Fractal I have an urge to back it off as I know in the real world it would be hard on the tubes. LOL

[@AndGuitar]

More chewy and more greasy, great descriptions! 😀 Excellent post!

[@mookytc]

Love that JCM800. Happy holidays Jason!

[@hrsey71]

fantastic video! i learn so much watching your channel

[@HeadfirstAmps]

Great to hear!

[@hrsey71]

@@HeadfirstAmps do you take any technical questions? im having a hell of a time troubleshooting a red plating tube issue

[@HeadfirstAmps]

Sure, feel free to post over at my DIY builders group. It’s mainly for HF pcb builders, but we are getting more general stuff in there now which is fine.

[@hrsey71]

@@HeadfirstAmps got a link?

[@HeadfirstAmps]

Search for Headfirst DIY Builders group on FB

[@deviantmultimedia9497]

That was a good one, mate. Thanks for doing that. I was surprised to see the ringing on the recovery.

[@qddk9545]

Super interesting. Food for thought: You can get SCR, I think up to 50uF 630v plastic caps now at reasonable prices. They will last forever, and maybe sound better, you can get away with half the caps on a 100w amp because of the higher voltage, but they don´t come in double cans. Some amp gurus think it is better to distribute the caps, instead of placing them at the start of the amp.

[@stephenchow5161]

i love your easy to understand explanations!

[@HeadfirstAmps]

Tnx Steve, will try to keep them coming!

[@curtisprice9806]

I liked the 16uf's sound! Thanks for the test and sound demo! I would have thought that 50 or higher would have been better, but my ears liked the 16's better!

[@HeadfirstAmps]

Thanks for watching!

[@myuncle2]

Jason when you don't use any delay you sound much better!

[@HeadfirstAmps]

Good feedback and duly noted!

[@buzzedalldrink9131]

@@HeadfirstAmps you don’t need anything! Your tone and playing are awesome as is!!

[@HeadfirstAmps]

Thank you!

[@cheyneluke9330]

Great video Jason thanks for dong this. Really informative!

[@johnmarshall3903]

More excellent information, thanks Jason.

[@HeadfirstAmps]

Cheers John!

[@jasonstone1046]

Informative as always mate and diggin' the riffs xD

[@HeadfirstAmps]

Tnx Jas!

[@ModulusAmps]

Good video Jason. For fun change the output tap to change the primary impedance and see how that effects screen current. IMO this is a big reason trainwrecks are set up as they are.

[@HeadfirstAmps]

Cool idea!

[@gil_evans]

Interesting!

[@batpigguitar]

That sounds bloody amazing

[@HeadfirstAmps]

I’ve playing this amp all afternoon, loving it more and more set like this….I’m going to have to try it with the variac next!

[@Jonathan_Doe_]

50uf for me, the 16uf sounds gritty on the harmonic overtones (in a bad way) and seems to lose some chime and bass. It’s not doing that ‘sag’ thing a JTM45 does when you hit a big chord either, I think that’s more down to the valve rectifier than the filter caps. Valve vs. Solid rectifiers would be an interesting video but slightly trickier to do (unless you just borrow a mess with a rectifier type switch)

[@eyedunno8462]

Okay

[@BadToad1963]

What you're calling "recovery ripple" I would call "ringing". That's the first thing I noticed when you put that up.

[@ericharbauer]

Fantastic video! TY!

[@rgbplumbinghilton]

The 50uf definitely makes it slightly fatter tonally. I can hear that through my monitors.

[@musicrepairs]

i think i prefer the tightness of the 50uf. But its splinting hairs

[@RunningMan40510]

First thought at beginning of the video is (this amp sounds very closed and not breathing well.

[@PhuketMyMac]

Excellent! Really, really great content. Invaluable! Please keep these theory videos coming. Do you have a Patreon page?

[@HeadfirstAmps]

Thanks! Not yet, but I’d like to get something started this year!

[@chrisstevens4680]

Question, how does this ringing affect the audio in practice?

[@GTguitarworks]

Will Old filter caps make the amp sound like it has a lot more fuzzy distortion and cause the bass to be muddy? No worries.. I answered my own question lol, I replaced the filter caps today, including bias caps and the whole bottom end tightened up, Those 40 year old filter caps where making my amp sound like shit... Now it's tight and crisp...

[@HeadfirstAmps]

Good find!

[@dannbrauckmann1545]

So cool! I've read about folks saying they notice changes if they lower the filtering in the Preamp Filter caps, as well. I've got 2x 47uf caps in my preamp. Would I notice any difference if I dropped those to 32uf or even 16uf?

[@evitative462]

Beat that calculator like it owes you money.

[@HeadfirstAmps]

Ha ha ha 100%

[@robimiara7444]

hhh

[@conjering]

Filtering does matter.

[@pjoliver182]

I always felt the preamp filter cap had the biggest effect when dropping lower than 50uF

[@eyedunno8462]

It has as much effect as any other filtering node. But in this video he is going over Power tube distortion where Marshall's true tone lies. It can be unforgiving at higher volumes even modern metal, so it helps for the filtering to have just a bit of give. Typically 47uf or 50uf is about the sweet spot for tight and chewy, lower is a bit more 'carmelized,' more personality and give but doesn't hold up under higher gain. 100uf has been used in late 80s JCM800s as ultra-high gain players demanded ninja-like response from their Marshalls, but played anything slower than blitering pace and gain yielded a sterile, uninspiring tone A fair compromise today is for amps to have stiff mains/plates filtering of 100uf or more, 50uf for screens and PI, 22uf for preamp. Tight where it matters, forgiving where it matters

[@johnsfixton8857]

I have a 18 watt JCM800 from Tube Town . Solid state rectifier . It has 2 2x100uF and no B5 . I put 100 at mains and 50 at the rest , but the amp become noisy . So what about noise ? If you lowered the filtering .

[@HeadfirstAmps]

Check your grounding of the new caps, a bad ground will introduce bad 100 or 120hz hum

[@johnsfixton8857]

@@HeadfirstAmps 100hz with master at 0 and it goes up as master goes up .Not very loud , the fans of my desktop pc cover it . So i wonder about the diodes it has 2 UF5408 not 4 1N4007 . I don't think it is a ground . I will change the diodes .

[@HeadfirstAmps]

100hz with master on zero is usually, a bad power tube, bad mains filter grounding or bad mains filter cap…or failed choke.

[@pedrolourenco2707]

both sound good! ...what are the preamp mods on this one?

[@wakjob961]

50/50 for the 50 watters... 16/16 for the 100 watters. That's just my taste. But whatever, you and your amps sound really great man.

[@HeadfirstAmps]

Thanks. I think you are right on with the distinction that between the 50w and the 100’s. the 50’s are already a little sweeter.

[@stevelark9840]

Again, another facinating video. Did you notice any increased hum, going from 50uF to 16uF on the screens? Do you think there would be any adverse issues on going this route on a 2204? Thanks Jason!

[@HeadfirstAmps]

Thanks Steve. No noticeable increase in hum, the filtering through the rest of the preamp will see to that. I have 32+32 on the pres. Do try it on a 2204!

[@Spentelectrons]

I can't speak to the 800 or 2204 but in a fender vibro champ scratch built with a very small power transformer. Just enough transformer to do the job. It had quite a bit of hum but was drowned out by the signal to noise while playing. For an experiment I stuck a 220uf cap after the choke. It's tube rectified. The hum dramatically lowered. After some testing i concluded the additional capacitance raised the b+, increased the bias, lowered the sag, raised the rails decreasing distortion, and dramatically decreased hum. The b+ is reducing in ripple, sag spreading the rails, and cleaning up the sound a bit. It's all semantics to me though with knobs and pedals involved. I would 100% jumper in another cap and see how it sounds. No harm but the cost of a cap

[@KB_Joys]

I have 100 on the screens in my 2204. It goes 100, 100, 40, 20, 20. Ive also 47n PI coupling caps, a 470k resistor into a 500k master volume pot. One thing I've noticed with every 2204 ive built, variants etc, is there's an artifact that sounds like over distorting, or parasitic inductance, whatever but ive finally solved the problem with a 120pf cap on THE OTHER SIDE on the PI over the 220k bias junction.i also removed the 47pf over the 82k/100k input side of the driver. Have you ever tried it? Seriously try it, you'll thank me later☺️ Honestly its why i always felt like its why i always had to increase the negative FB with those amps. If you do the 120pf mod you'll find that you can run the 100k on the 4 ohm tap and it sounds like you think it should.

[@miseryparty3726]

Where exactly do i do this mod on the circuit?

[@KB_Joys]

@@miseryparty3726 my amp is a homebrew. But you'd have to tap into the power supply, where you'd want the filtering or extra filtering. I believe in some 800 amps they only have two 10 k resistors, and can have a too high voltage in the preamp. I like the way they did Kerry Kings amp where they take the power node for preamp to the power amp. From the way I wired my jcm800 clone I did a 50 watt version on the KK 2203. It allows you to use lower voltage to the preamp, but without making it sound too thin running from the screen section of the power supply.

[@miseryparty3726]

@@KB_Joys I meant where exactly did you put the 120pf cap? my negative feedback is pretty weak in my 800 lol

[@KB_Joys]

@@miseryparty3726 oh, I thought I was replying to something else, lol 😂 As far as that capacitor and where they normally place it across the plates, you can parallel it with one of the PI plate resistors. I don't remember which one off hand, but you can also place one across the the plates on the other side of the resistors. Also id like to point out that I noticed some instability with a few amps during the initial build process. Like after I got things together, ran tests, set bias, etc things seemed to just start holding together and had more stability if you will. I thought it was weird, idk maybe it was just me. Anyways I ended taking that capacitor out because things were stable.

[@wenzelli]

So which do you prefer, cranked 1987 or cranked 2204?

[@HeadfirstAmps]

Oh the 1987 cranked is very hard to beat, but the 2204 is more useable in a band context. The 1987 really needs the Fryette PS2 with it to make it band friendly. I prefer the 1987 though tone wise.

[@myuncle2]

@@HeadfirstAmps What is surprising is that for recording purposes, the JTM 1 watt (the little head with a nice 2x12 cab, not the combo) is hard to beat, yes you need to crank it but you don't need an expensive power attenuator. It gets as good as a cranked 1987, and even more chewy. Do you think it depends on the fact that it has no EQ knobs? Sometimes I think the less knobs the more clarity.

[@dwayneanderson7108]

Not as much as people seem to think...

[@jordanreinecke1378]

Is 100uf for screens and pi in a 50w with gz34 a good idea? To keep it tight and keep the rectifier sag

[@qddk9545]

Just be careful with the main filter cap, according to GZ34 datasheet, I can´t be larger than 60 uF. According to your question: It will give a very stiff supply, so you must decide if you want that.

[@RozsaAmplificationLLC]

Nice video Jason! I like the 50's more than the 16s. The 50s are more smooth and the 16s in the video are too "edgy"!! Cheers! Oh, where is the link for comparing chokes?? Didn't see that!

[@HeadfirstAmps]

Edgy is a good description! Perfect in fact. I like to play the edgy setup, it’s easier! Lol Chokes, interesting…one to add to the list!

[@RozsaAmplificationLLC]

@@HeadfirstAmps Oh. I thought you commented you had already made a video! My bad. Must have misunderstood. Yeah, edgy would be a personal trait. I like sizzle (ala Doug Aldrich), but not "edgy". I like the soft rounded in the power section and adjust the sag or tightness in the preamp section. Friedman sort does this. His power section is really fixed and in the preamp will run anything from 16, to 22 to 33 to 47 depending on what he's looking for. I find the Jose pre-TS more "edgy" also, but like that edginess...but again it's in the pre. I think you do also. Friedman....not so much and prefers the standard post-TS sound. One suggestion on the power section and when presenting it and explaining it....you utilize a standard Marshall (Friedman) setup where capacitors are in series and using balancing resistors across them. Would be much simpler to just use a capacitor and NO resistor as the other stages. Really, the only reason to do this is when the voltage is greater than the rating of the capacitor itself and in the old days running 540-800V. When running 500V or less...just run as I said and a capacitor and no resistor and on the plate and screen and eliminate components and cost and those caps are not cheap and $0.15 resistor!!! Unless high voltage, I don't use the extra caps and resistors on ANY of my amps. Wasteful and costly...and easier to explain (instead of saying, well their are the same value and half value or having to calculate it out)!!! :)

[@HeadfirstAmps]

Well, only all about chokes and comparing to a static resistor. I have done that. But I haven’t done a A/B compare of different choke specs. This one: kzbin.info/www/bejne/r4LNlWBvg55meZY

[@RozsaAmplificationLLC]

@@HeadfirstAmps k. static v. choke. I'll check it out. Thanks for the link!

[@cgavin1]

Always got to bear in mind that high wattage amps were designed to sound best cranked. A lot of years later we find that the original specs don't deliver the sound and response we want at the considerably lower volumes we want to hear them at. Increasingly this is true of speakers too. People go higher and higher power handling, thicker more rigid cones, bigger magnets etc. The venerable pulsonic Celestions were paper thin and flapped around like your grandad at a hen do! A good analogy is the sports motorcycle industry in the 00's. They kept producing these monster horsepower road bikes that did absolutely nothing below 70mph/10, 000 rpm. People didn't understand why the powerband was so "unusable". They just weren't as much fun as the torque monsters. The answer was because they were not designed for the road...

[@kasimirokaz3775]

Do these filter caps also give white noise sounds and some crackles when they begin to go bad? Some days my Marshall does that and some days it doesn't. I'm no tech or expert thus the reason for my question.

[@curtisprice9806]

Yes. Lots of crackling like you hear when frying bacon. Also ghost notes following the notes you play. Filtering caps make a huge difference. My old Marshall superleads need their electrolytics replaced because of hum and response too slow

[@GTguitarworks]

Would you recommend going this low / 16Uf on a JCM 800? I was thinking 32uf would be better to retain the correct feel of an 800? because they're not supposed to be as spongy as a Plexi, Maybe? but wouldn't mind a tad bid of sag on my 1983 800

[@HeadfirstAmps]

For a JCM800 2203/4 I'd have 50uF.

[@GTguitarworks]

@@HeadfirstAmps 0K, That's what I will put, Thanks!!!

[@kristerpanzerfaust]

Nice video! Would the feel of playing 50 vs 16uf be comparable to playing ss vs tube rectifier? Always wanting to have a jtm45 only because of that saggy feel everyone is talking about , so if i would put 16 + 35uf with a switch between make it a jtm+jmp in the same amp?

[@HeadfirstAmps]

Sure, maybe 16 & 50 would be more noticeable (on the screens). And you have to be cranking the amp to really feel the difference.

[@roncarter2188]

The info and knowledge presented here is great....but I'm listening because of his accent. 😅

[@HeadfirstAmps]

Nah mate, you’re the one with the accent.

[@cap217cap217]

Did you try this test in the preamp cap?

[@HeadfirstAmps]

Not in this amp, but yeah I have. Not as a bigger impact. Power section filtering makes a bigger difference.

[@BracliCreations]

Why aren't full wave rectifiers used in Marshall power supplies, isn't full wave have less ripple than two phase? Pls educate the diode illiterate

[@HeadfirstAmps]

They are in the 100w amps. With two phase, like in the 50w's, you still get the same ripple factor ( 2 x mains frequency). It just goes about it a different way.

[@BracliCreations]

@@HeadfirstAmps Actually I've observed that most tube amp circuits use two phase rather than full wave bridge rectos. I wish had studied EE so that I may better understand 60 year old analog circuits lol. Thx again for creating this content!

[@qddk9545]

@@BracliCreations It is because of the transformer voltages.

[@rjschmirler3794]

So…. A 1987 circuit is different from a JCM800 (2204). Other than that this guy is alright..

[@HeadfirstAmps]

It’s not a 2204. It’s a 1987 circuit. JCM800’s were a series, not a circuit. There were 1959 and 1987 circuit JCM800’s.

[@qddk9545]

You got an answer, but yes the 1987 is a simpler 4 input amp with no serial coupling of the gain stages and a MV. To complicate things the first few 2203 (100w) and 2204 (50w) did not have the serial coupling and MV. Because of the you have to play the 1987 hard to get your distortion.