Tiny Terror: 216 uF capacitance for just 15 Watts?

[Hardtailed]

I know the preamp caps don't phone home any current.

The front wheels serve a different purpose than the ones in the back, but we still call them 18-wheelers, right? It's a seat of the pants comparison, no more than that.

Power supplies are very similar, so when I said it has 216uF total capacitance, if you go cap by cap you'll see that most of the time the proportion holds. If it were a JCM800 and kept the proportion per watt, it'd have 1440 uF, a senseless measure to illustrate that this 15 Watter has bigger sag fighting ambitions than Laney's, for example.

There is zero point in adding the preamp capacitors and comparing that to the power of the amp. They are isolated by big decoupling resistors which only see a few mA go through.

And also: why does it matter anyway? Are you afraid that Orange robbed their clients of some super special magic tone by using too much capacitance?
The filtering is just one ingredient in a big recipe.

Also, remember that the rectifier feeds the amp in very short bursts (the diodes only conducts when the voltage is higher on the anode side than on the cathode side). The capacitors are the sole source of current for the majority of the time. Add to that that the voltage in the capacitors will drop as SOON as you start discharging them (drawing current). A bigger capacitor will simply discharge less quickly (less voltage drop for a given charge).

No matter how big the cap, there will always be some sag.

There is no such thing as too much capacitance, it depends on the "recipe".

Stop agonizing over this

[jmaf]

Take the ubiquitous 2203 (single channel 100W) for example:
http://www.drtube.com/schematics/marshall/2203u.gif

Notice that after the bridge rectifier it hits 2 100uF capacitors connected in series.

Is that a JCM 800?

By your "math" that would amount to 200uF

Did I add capacitors wrong somewhere in my examples? I didn't notice, it'd be possible considering it's friday and I'm tired as heck.

[jmaf]

There is zero point in adding the preamp capacitors and comparing that to the power of the amp.
[ ... ]
Stop agonizing over this

OK, thanks for your feedback, appreciate it.

[Hardtailed]

Is that a JCM 800?

When most guitarist say "JCM800", they usually mean the 2203/2204 (100W and 50W version respectively) circuit. It's the one with only one channel, no reverb, no feature. However, that circuit was available in the JMP years. JCM800 barely refers to a cosmetic make over (the biggest difference being the full width chassis). The non-master volume circuits (namely the 1959 and 1987 circuits, which is what most people mean when saying "Plexi", these do not refer to years but model numbers) were still available in the early JCM800 years.

Throughout the years, other models were released in the JCM800 line, namely the channel switching 2210/2205 versions. Then others with reverb, or the famous "Jubilee" versions which had diode clipping for extra gain. As far as I know, none of those ever had more than 100uF of filtering on the first node (the one that counts).

There is a lot of interesting information about the history of Marshall on the internet if you are curious.

Did I add capacitors wrong somewhere in my examples? I didn't notice, it'd be possible considering it's friday and I'm tired as heck.

Well the point here is you can't just add capacitor values and try to read something out of it. On top of the fact that preamp caps don't really matter, sometimes the caps are in series (2 100uF in series totals 50uF).
This is done to get better voltage handling (the voltage handling adds up when capacitors are serial connected).

Take the Soldano SLO, the main filter caps are 4 220uF units. But the total capacitance is... 220uF! Add to that that it uses small caps in the preamp and you'll conclude that total capacitance is about on par with a 100W JCM800, yet it is MUCH stiffer because of the 4X higher filtering on the first node.

[jmaf]

were still available in the early JCM800 years.

Never seen a JCM 800 with cans in series. I just asked because you JMP'd all over me for using the JCM800 as example.

Anyway, thanks again for that bit of Marshall lore.

[fyl]

The "power" filtering is the one fed straight from the rectifiers and going to the output transformer center tap, and its value can and will influence perceived sound.

The first filtering node (plates) is the most important but the second node (screens) is nearly as important - particularly with true pentodes where screens can draw sizeable amounts of current.

Dimensioning the first cap is quite easy: just shoot for less than 10% ripple voltage. Contrary to intuition, ripple isn't linked to supply voltage but to supply current - you can have very high levels of ripple in poorly filtered DC heating circuits and very low levels in high voltage/low current supplies such as those used for preamp stages.

Vpp = I / (2 * F * C) for full wave rectos, I = max current, F = mains frequency, C = capacitance

For example a 47 µF cap gives app. 35 V ripple @ 200 mA, OK for most supplies.

This gives a minimum, you can then play with values in order to "tune" the node.

Preamp capacitors are hum (and motorboating) killers , the larger the merrier in general terms.

They also cut some HF noise.

Sizing the resistors and caps is also pretty easy: aim for a low time constant, ideally below 1 Hz.

Fc = 1 / (2 * Pi * R * C)

Say R = 10K ohms, C = 22 µF, Fc = 0,72 Hz

Then play with values in order to obtain different time constants for every node, you'll be pretty sure than the amp will never motorboat. Or you can rely on component tolerances - elcos are ±20% parts at best - and use the same RC values everywhere.

[jmaf]

At 90mA (Edited: PEAK current)current for both EL84's we have:

Vpp = 90mA / (2 * 60 * 100uF)

Vpp = 7.5 VAC ripple

Unless I got the algebra wrong, if shooting for 10% ripple we could power up 4 dimed Tiny Terrors with a single of these power supplies. Orange went for 2.3% ripple during maximum PEAK current here.

[fyl]

At 90mA (Edited: PEAK current)current for both EL84's we have:

The TT is in deep AB1, with app. 90 mA plates / 20 mA screens, for 110 mA max. Plus a few mA's for the small bottles and the PS glue. Say 120 mA max.

[jmaf]

The TT is in deep AB1, with app. 90 mA plates / 20 mA screens, for 110 mA max. Plus a few mA's for the small bottles and the PS glue. Say 120 mA max.

You're right indeed, it ain't just the plates dangling on that reservoir(TGI friday). 10 VAC PP ripple it is then, 3% ripple.

Sidenote: It's advertised as a pubrebred Class A.
Orange Amps | Ranges & Products | Tiny Terror Head

[Hardtailed]

Sidenote: It's advertised as a pubrebred Class A.
Orange Amps | Ranges & Products | Tiny Terror Head

Yep, Vox has been getting by with calling their AC15/AC30 "Class A" for so long that nobody questions it anymore.
The common marketing approach is to consider anything cathode-biased as Class A.

The funny thing is how people assume Class A is automatically better (probably because it implies it is the best class, while it simply was the first one to be invented)

My favorite is when people think a "Class A watt" is louder than a regular watt.
AC30s are not loud because of their class, they're loud because 30W is loud in the tube world

End of rant

[jmaf]

My favorite is when people think a "Class A watt" is louder than a regular watt.

Funny you should mention that.

I got into an issue earlier this week with some knowitall catalog addict who was trying to make exactly that case.

He said power attenuators for Class A amps should be more sturdy than attenuators for Class AB.

Class A probably sells better and I have a strange hunch that the British are to blame

[Hardtailed]

Funny you should mention that.

I got into an issue earlier this week with some knowitall catalog addict who was trying to make exactly that case.

He said power attenuators for Class A amps should be more sturdy than attenuators for Class AB.

His thinking might be based on the fact that true Class A amplifiers dissipate a lot of power at idle. However, what he fails to understand is that this idle current is pure (or almost pure) DC, none of it makes it through the output transformer (which only transforms the AC component)

True Class A amplifiers are very uncommon in the guitar world, save for the few single-ended amps like the Epiphone Valve Jr. And that should be a proof that it's not automatically better.

[J M Fahey]

Hi Fyl. Fully agree with your math, just didn't want to make the post very long.

[guitician]

This amp would crush that tiny terror. Look at it's PSU!
http://music-electronics-forum.com/t5328/

[Steve Conner]

Wow, 25uF total.

I think the deal is just that electrolytics are so much smaller and cheaper per joule of stored energy than they were back in those days, thanks to the invention of etched foil or whatever.

Also we now have SS rectifiers that don't care how much capacitance you hang off them. (the transformer can get upset with too much, but that's another story)

So amp designers now use more capacitance just because they can.

Also, the solid-state era raised people's expectations with respect to hum level. Transistor amps use so much NFB that they can cancel out all of the power supply hum even with modest-sized filter caps. People who grew up with them expect their tube amps to be hum-free, too.