Power supply capacitors

[KT66]

Can anyone tell me if there is such a thing as too much filtering in a guitar amp power supply and if so why it is bad? I built a 50W plexi with 5 filtering stages, the first two are 2 X 220Uf in series and the last 3 are 20Uf, but I am thinking about changing the second stage to 20Uf because I read somewhere ( Dumble discussion I think) that 2 beefy filtering stages should be used on a 100W amp but only one for a 50W. Also, on a somewhat related note, I was reading in TUT that decoupling the power supply so that there is a filter for every single gain stage helps reduce noise, have any of you tried this?

[pila]

I was reading on another forum that the rectifier tube data tells you the max size for electrolitics. Don' remember if it's only for the first stage after the tube, or are you using a SS rectifier?

[KT66]

You're correct that you can only go so big with a tube recto or you get motorboating. My amp is solid state. I'm really looking to find out if there is some reason tone-wise or performance wise that you shouldn't have 2 big filtering stages in a 50W amp. The amp works and has a tight bottom end which is what I was going for, but my next build is going to be an ODS so I have been studying the schematics and reading everything in the Dumble forum and it seems to me that if HAD had one big cap and then all 20s then there must be a reason other than economy because he didn't seem to care about that from what I can tell.

[unklmickey]

...The amp works and has a tight bottom end which is what I was going for,...

i think that's part of your answer.

if you want an amp to be "stiffer", and less prone to sag, larger caps in the first 2 stages will help (as will SS rectifiers).

since there is very little "resistance" in SS rectifiers, it is usually necessary to increase the size of the cap in the first section, to keep the same level of filtering.

if you want more sag, smaller caps and tube rectifiers are the way to go.


cheers,

unk

[UR12]

I'm really looking to find out if there is some reason tone-wise or performance wise that you shouldn't have 2 big filtering stages in a 50W amp. The amp works and has a tight bottom end which is what I was going for, but my next build is going to be an ODS so I have been studying the schematics and reading everything in the Dumble forum and it seems to me that if HAD had one big cap and then all 20s then there must be a reason other than economy because he didn't seem to care about that from what I can tell.

Filter caps not only filter out 120hz or 60hz power supply ripple but also act as small batteries that can instantly supply current for short bursts. The larger the caps the longer the bursts. Most of the current flow in an amp goes through the power tube's cathodes and plates. This is the first filter stage and you can use a larger value cap here (with SS diodes of course) to increase the stiffness or bass response of the amp. Bass notes require more power to produce than high notes. That's why if you have an amp with a poorly designed or weak power supply you get the bass notes sounding mushy or "farting out"

The next filter stage is for the screens, as they are the next biggest user of power. (probably about 1/10 or less the power used by the first stage though) Since the demand for current isn't as much as the power amp first stage the caps can be smaller.

The preamp stages draw the least amount of current. If you go over about 3 ma for a 12ax7 you would be pushing it a bunch but for just for example lets say that they will be pulling 10ma of current. Compared to the 30 to 150 ma that the power tubes are drawing they are using only a small amount and don't need much in the way of caps to supply them with a bunch of current. Since the power stages have already filtered the "hum" twice already then less capacitance is needed on the preamp tube because they will have less hum to filter. Why spend money on huge filter caps for the preamp stages when it isn't needed?

So gettin back to your original question about tone, Bigger filter caps in the first 2 stages stiffen up the PS and produce better low notes so there is some effect on the tone of the amp.

[sliberty]

Hey Dana - thanks for taking the time to explain that so clearly. That was a really valuable learning experience for me.

[Spinoo]

One must add that with a capacitor input filtering the transformer has to 'fill' at least the first capacitor (depends if you're using a choke to smooth the screens' supply) in a very small time, which leads to very high current spikes.
These spikes cause the power transformer to heat as magnetic saturation happens, a friend just burned one using 69µFs for a preamp.

[UR12]

One must add that with a capacitor input filtering the transformer has to 'fill' at least the first capacitor (depends if you're using a choke to smooth the screens' supply) in a very small time, which leads to very high current spikes.
These spikes cause the power transformer to heat as magnetic saturation happens, a friend just burned one using 69µFs for a preamp.

Very good point! I might also add that the charging current goes through the diode rectos also and you need to make sure that you size them accordingly or you will be exceeding the current rating on the diodes. There is a lot to take into consideration when designing a power supply and an error in judgement can cause SERIOUS problems. Some of the DC heater supplys use a 10000 uf cap as the filter. It is good practice to use a 10 - 25 amp bridge just because the current to charge these large caps will fry a 1n4007 rated at 1 amp and we are only talking 6 volts.

[KT66]

Spinoo, I don't think that the momentary inrush of current with larger than normal preamp caps should cause a PT to fail, there probably was another reason why it did. Having large filters on the preamp stages is unneccessary though because after the second cap in the chain the rectified voltage is basically completely smoothed out. If you have a scope you can check this out. I think that the main reason for having more than 3 caps in the supply is to decouple the stages of the amp. Some amps feed a whole bunch of the preamp plates from the last filter. Look at a Supro S6424, all 3 of the preamp tubes are filtered by the 3rd and final cap. Or check out how many tubes are powered after the final PS cap in an Ampeg B12XT. Anyhow, I think I'm gonna take out the second large filtering stage in my plexi to make room in the doghouse for an extra cap so that I can add a tube recto switch, 'cause I do like the sag and because I have an unused 5V tap, but mostly because I love having lots and lots of switches all over my amps so that they look like a mad scientist with a switch obsesion constructed them.

[unklmickey]

hi Ryan,

i was a bit puzzled by Spinoo's anectdote when i first read it too.

i wonder if he might have been talking about a separate preamp, not the preamp section of a complete amp?

in any case, i doubt even with a big cap on the output of the rectifer, the surge could cause the PT to fail................BUT,

it could start a cascade, by causing SS rectifiers to fail.

if they short (not rare at all) when they fail, that could cause the PT to fail.

...but mostly because I love having lots and lots of switches all over my amps so that they look like a mad scientist with a switch obsesion constructed them.

you are one sick dude!

(i'm told i was separated from my twin, at birth. what's your birthday?) [img:29:27]http://s21.photobucket.com/albums/b269/ ... llinme.gif[/img]


cheers,

unk

[pila]

Hey Ryan, you should be piloting a corporate jet, tons of switches!! I used to have to fix all that stuff!~!

Pila

[Spinoo]

Well to make sure you power supply is correctly sized you could use Duncan's PSU Designer.

I was talking about a standalone preamp - nothing in common with a Trainwreck - my friend used a PT with the secondary rated @ 80mA and after simulation his cap choices caused current peaks @ 400mA. He used no choke at all to smooth this out, and had about 10 triodes to feed.
With a capacitor input filter such current peaks are to be expected, but his caps were so big that the PT blew

[KT66]

Well, I removed the second filter and replaced it with a 20U, so now my filter string is 110U, 22U, 22U, 22U, 22U instead of 110, 110, 22, 22, 22. I found out that the answer to my original question, and that is, yes, you can have too much filtering. After I replaced it the humming that I was getting from the amp was drastically reduced. I decided to try this because of something I read over at Paul Rubys site. Now, I still have hum, which I think is from the filaments, or maybe being induced into the OT from the PT, so I decided to try elevating the center tap of the tube heaters. Instead of reducing the hum it introduced a buzz into the amp. Yes, I have the DC to elevate it filtered enough, but I was wondering if any of you have any thoughts on this matter. Also, my ground scheme is pretty solid, I used a buss bar andhooked all of the preamp grounds and filters to it and grounded it at the input end of the amp. The OT secondary common, power tube cathodes, HV center tap are grounded near the PT.

[unklmickey]

...so I decided to try elevating the center tap of the tube heaters. Instead of reducing the hum it introduced a buzz into the amp. Yes, I have the DC to elevate it filtered enough,...

in theory, if the heaters are ALWAYS more positive than the cathodes, the DC ripple shouldn't be an issue.
you don't want to get too carried away. (+50 to 70v seems about right.)
if you exceed the maximum heater-to-cathode voltage (100v for a 12AX7), you'll have other problems.


Ryan, i'm surprised about the 110uF for the screen grid supply causing more hum than a 20uF. are you certain one of those 220uF caps wasn't open?

i've heard of very large electrolytics being poor at filtering HF noise, because of the inductance inherent in their design, but we're talking 60/120hz here.

[Spinoo]

Hi, you should try to identify your noise : is it 60Hz or 120Hz ? Then you'll know if ot comes from the high voltage rail or from a magnetic coupling or heaters' supply.