Date: Fri, 2 Jul 1999 17:09:32 EDT
Subject: Re: paralling lytics
X-Mailer: AOL 3.0 16-bit for Windows sub 38
In a message dated 99-07-02 14:44:04 EDT, you write:
So the actual question is really, “what do you know about paralleling lytics with plastic caps, and how do you avoid setting up an unstable, resonant power supply?” Is it merely enough to throw a scope across the B+ rail and put the amp through its paces in order to determine what’s going on? In my case, would I be able to determine the optimum value of C by monitoring for signal voltage on the B+ rail–ie. keep adding C until I get a nice, stable voltage under full load, square wave conditions?
I may end up posing these questions in a more public forum, but seeing as you’ve provided so many answers to my questions in the past (and it’s good advice as well) I figured I’d give you a shot at it first.
As always, I greatly appreciate any wisdom you can chuck my direction.
The resonant condition, although possible to set up, is widely overblown.
Modern (i.e., built within the last 5-10 years) aluminum electrolytics have a very low “ESR” that holds to about 50 kHz before becoming a problem. This means that the cap, by itself is usually good enough. However, the older ones start losing their effectiveness above a couple kHz. Sounds like you’ve got some of the older ones. These have “relatively” high ESR along with relative high inductance. The high ESR helps you in this case….. you can parallel smaller but more effective capacitors without any odd resonances showing up, simply due to the low effective “Q”.
To test the condition, short of having a network analyzer, the test you propose is quite adequate. Look at the effective ripple, particularly noting overshooting or undershooting of the square wave edges. One additional test (watch, cause its a little tricky)….. use a capacitive load on the amp (this causes a phase angle in v-i supplied by the PS, but you also have to watch you don’t overdissipate the tubes either). Instead of 8 ohm resistive, use 5.6 resistive + 5.6 ohms reactive (at the fundamental) for an effective 8 ohms at 45 degrees phase angle; this simulates the worst case load reasonably well. If your PS doesn’t wobble under both conditions (resistive and R+X load), you’re pretty safe. I’d probably use something like a square wave in the 2-5kHz range, since this will allow significant harmonics to pass as well.
What you may find is you can successfully parallel “plastic” directly across those BFC’s.