Capacitors ripple current question

[catalin gramada]

How can be determined/estimated in a very simple way the ripple current performance for a filter capacitor, please ? I have to replace screen grids filter caps in series battery into a Sound City PA 100 and I just have two small FT rated at 0.2 A ripple current.This is ridiculous small relative to originals (big cans rated at half an Amp).Can I use it safe, please? Thanks

[J M Fahey]

How can be determined/estimated in a very simple way the ripple current performance for a filter capacitor, please ? I have to replace screen grids filter caps in series battery into a Sound City PA 100 and I just have two small FT rated at 0.2 A ripple current.This is ridiculous small relative to originals.Can I use it safe, please? Thanks

You ask too much offering very little data
two small FT rated at 0.2 A ripple current.This is ridiculous small relative to originals

What does "small" mean?

What does "originals" mean?

Canīt find those units either in my Capacitance meter nor in any equations

I can only find "microFarads" , "current", "voltage", etc.

And searching datasheets stupid Google asks for "Brand" and "model/type".
Computers are dumb.

Please show the circuit they are intended to work in, with voltage values.

[catalin gramada]

Did not found a schematic for this model. I have to draw one by myself and come in back

[Chuck H]

We really only need to compare the voltage rating and uF value of the original cap to the new ones you want to use.

[nevetslab]

Ripple Current Characteristics in Aluminum Electrolytic Capacitors

How can be determined/estimated in a very simple way the ripple current performance for a filter capacitor, please ? I have to replace screen grids filter caps in series battery into a Sound City PA 100 and I just have two small FT rated at 0.2 A ripple current.This is ridiculous small relative to originals (big cans rated at half an Amp).Can I use it safe, please? Thanks

I have extracted the three pages pertaining to how ripple current is derived from a technical paper from Nichicon.

Nichicon Ripple Current in Electrolytic Caps.pdf

[Chuck H]

I have extracted the three pages pertaining to how ripple current is derived from a technical paper from Nichicon.

[ATTACH]50975[/ATTACH]

Indeed! I've never been more intimidated. How much easier to choose a cap of appropriate ratings for the circuit with a decent impedance spec and then measure the ripple current.?.

[catalin gramada]

At iddling should be a simple task. Have no idea how to do the measurement with a multimeter into dynamic conditions when think dramatically changing load currents generate also pulsating currents into decoupling caps as ripple...

[Reader]

I believe that the reality is that normally available standard commercial quality electrolytic capacitors will be OK with respect to ripple current capabilities in the typical musical instrument power supply applications. My understanding is that ripple current becomes an issue when designing high efficiency switching supplies and conventional supplies that run at high ripple values due to high current draw. In the case of high current draw the circuit being fed may be OK with the high ripple. In musical instrument amplifier circuits the first filter stage ripple is usually less than 4% and then much much lower for the preamp stages. Therefore, catalin gramada, I do not believe that you do not need to worry about the ripple current specifications of the caps you want for your Sound City PA 100.

[Dave H]

How can be determined/estimated in a very simple way the ripple current performance for a filter capacitor, please ? I have to replace screen grids filter caps in series battery into a Sound City PA 100 and I just have two small FT rated at 0.2 A ripple current.

You could use PUD2 as below. I've simulated it using the plate and screen currents of a 4 x EL34 amp running at 100W using the circuit of a typical Sound City amp. It gives the ripple current of the plate supply cap (IC1) as 800mA and the screen supply cap (IC2) as 100mA at 50Hz mains frequency.

Soundcity_100b.pdf

[nevetslab]

Indeed! I've never been more intimidated. How much easier to choose a cap of appropriate ratings for the circuit with a decent impedance spec and then measure the ripple current.?.

That's probably why they put these three pages in the white paper 22 pages back from the start, where Nichicon tells you all you never thought to ask about, and take you thru all the stuff involved in producing the parts that we take for granted, and deal with them from the standpoint of some vender's parts that ALWAYS fail early. Size matters, in parts that we put in the highest ripple current area in our power supplies....that first stage. We're pulling all of our smoothed 'DC current from there. Temp ratings, ESR, published ripple current ratings, Dissipation Factor and, of course, physical size AND Cost/availability come to mind when I'm either replacing some parts, or specifying a part in a new design.

I too have never stopped to figure it all out, and leave it to the companies that do. Makes my head swim too, seeing all of what goes in to produce a generality ripple current rating chart that some data sheets carry, while other's don't. As our OP observed, the original filter cap in a vintage amp was very impressively large, then you're looking at some far smaller parts put in as replacements....what's going on here? Advancements in capacitor foil technology over the years for certain, and other refinements in packaging that achieve comparable internal temp rise from the applied ripple current have made our larger parts of old a whole lot smaller.

If size and cost are not a huge factor, I'll always select a larger part that has better DF, lower ESR, higher temp rating (105 deg C vs 85 deg C) & higher ripple current rating, just so I can be assured I've equaled or exceeded the quality of what I've replaced.

[J M Fahey]

Not reading the Nichicon paper which must certainly be very accurate and full of Math, simple thinking tells me that peak ripple current through a reservoir cap is what they have to supply under dynamic conditions, beyond idle current supplied through screen resistors.

So if, say, screens take 10mA at idle and 50mA peak when fully overdriven, then delta I , the instantaneous current variation is 40mA, and that is the ripple current.

Please repeat this with actual current values, and you donīt need any schematic for that, just read tube datasheets and multiply by how many tubes you have.

Without looking (hint, do your homework ) , I guess even 0.2A must be ample.
Doubly so if those "small" FT go in parallel.

By the way, you didnīt post model and ratings yet
You expect us to guess without data?

How many uF is "small"?

[J M Fahey]

and the screen supply cap (IC2) as 100mA

[ATTACH=CONFIG]50976[/ATTACH]

Thanks Dave, .
That said, I think C1 ripple current value must be right, since you load it with what 2 power tubes take (the 4 of them never work at the same time) but C2 ripple current must be quite lower, I doubt screen voltage sinks down to 0V or Ground at any time.
So voltage across R2 is way less than 400 something Volts so reducing current demand accordingly.

IF those 1k resistors ever saw 400V across them they would dissipate 80W each (160W at 50% duty cycle) which is obviously not the case.

[Helmholtz]

At iddling should be a simple task. Have no idea how to do the measurement with a multimeter into dynamic conditions when think dramatically changing load currents generate also pulsating currents into decoupling caps as ripple...

Ripple currents produce thermal power along with the ESR. This causes an internal temperature rise which in turn affects the cap's lifetime. Without using actual ESR and RMS ripple current values, the ripple load can be evaluated via the increase of the cap's case temperature (Tc). For cap ambient temperatures up to 60°C, a Tc increase of 10°C caused by the ripple current is a conservative limit for most E-cap types, meaning that no detrimental effect on lifetime is to be expected. Of course the rated max. Tc value of the cap must not be exceeded.

Regarding guitar amps, it seems completely sufficient to make sure that the cap's case doesn't get hotter than its specified max. temperature at prolonged full power operation and max. ambient temperature, as the case temperature includes ripple heating anyway. If you really care.

Bigger caps have a larger radiating surface and will show less ripple heating. Also lower ESR types produce less ripple heating.

[Tony Bones]

I took the PSUD simulation a little further to look at the steady state ripple current. I read J M Fahey's comments regarding the definition of ripple current. His takeaway, I think, is there is no such thing as steady state. But it's easier to use the industry definition and forge forward, even if we get the wrong answer! Joking aside, it's possible to setup PSUD to put a step function in the loads, so one could explore Jim's ideas.

Anyway, in PSUD, you can plot any of the variables over time by checking the box on the left. Here's essentially the same circuit that Dave H sim'ed. I think that 300mA to the plates is over estimated, but this gives a worst case answer. Anyone that wants to know what happens when you change that should download PSUD and try it themselves.

Note that the numbers in the table include the startup, which is horrifying.



If we change it to start reporting after a 1 second delay, then we can see what's happening after startup, which is more realistic.



Here we can see that the the ripple current through that cap is less than 60mA.

[Dave H]

I think C1 ripple current value must be right, since you load it with what 2 power tubes take (the 4 of them never work at the same time) but C2 ripple current must be quite lower, I doubt screen voltage sinks down to 0V or Ground at any time.
So voltage across R2 is way less than 400 something Volts so reducing current demand accordingly.

Did I get it wrong? nothing new there then

I just used double the data sheet values of plate and screen current for 2 x EL34 at 50W. That's (2 x 106) x 2 plate and (2 x 21) x 2 screen. R2 isn't a screen resistor. PSD2 wouldn't let me put a current sink there to represent the total screen + preamp current so I had to use a resistor.

EDIT: Well what do you know? I see Tony managed to put a current sink where I had R2. It's easy when you know how (I do now)