Welcome, Guest. Please login or register.

October 24, 2020, 10:16:25 PM

Login with username, password and session length

Recent Posts

 

Author Topic: Quasi-n00b LM1875 problems  (Read 11064 times)

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Quasi-n00b LM1875 problems
« Reply #15 on: October 06, 2013, 07:48:22 AM »
Quote from: MrStab
i see both that and the original TDA2030 datasheet circuit i was using have the 100uF caps to ground - some gut feeling tells me that that would be better?

erum ... "better" than what?  Where do you have them connected at the moment?

Whatever else there is in power supply filtering, it is important that bypass caps actually bypass the amplifier, and this is best (can only be?) done with the bypass caps in close proximity to the amp in question, with the shortest possible path.  In the case of an amp that has a known high frequency instability issue the parallel 0.1uF are quite important, and should also be situated as close to the chip as possible.
If you say theory and practice don't agree you haven't applied enough theory.

MrStab

  • Chipper
  • *
  • Posts: 9
  • Chip Points: 0
    • View Profile
Re: Quasi-n00b LM1875 problems
« Reply #16 on: October 06, 2013, 12:45:13 PM »
i (think i) elaborated briefly earlier in the thread, but my posts are convoluted: in the TDA2030 schematic i'd originally used, the 100uF caps were dotted, as if to imply they were "optional", plus they weren't in the vero layout and another one i found on the web, therefore i omitted them but had considered putting them back in.

so is the proximity to the chip required for reasons similar to how i'd decouple, say, an LFO in a different application? or is there a different principle at work here? trying to learn as much as i can as i go

will work on a new layout tonight and hopefully start work over the next coupla days. after much paranoid quintuple-checking of the LM1875 datasheet schematic, ofc.

cheers
« Last Edit: October 06, 2013, 12:49:34 PM by MrStab »

phatt

  • Legendary
  • ******
  • Posts: 2141
  • Chip Points: 261
    • View Profile
Re: Quasi-n00b LM1875 problems
« Reply #17 on: October 06, 2013, 09:42:02 PM »
Still confused as to what you mean exactly,,,,pics or this will go on forever and just add confusion.

Word of advice,,, draw a schematic of the psu and no I don't mean copy and paste :grr I mean look at every wire in your build and draw it on paper or paint if you like then post it. Even better pics of the actual circuit as well.
Don't be shy we won't beat you up about rough wiring,, we will just guide you through the rough spots.

The idea of transferring what you build back to a schematic is you actually LEARN something very useful and get you brain connected to working out how to translate schematics to the real world and back an forth thereby double checking what you actually have built.
Phil.

Roly

  • Legendary
  • ******
  • Posts: 2184
  • Chip Points: 288
    • View Profile
    • Australian Valve Amps
Re: Quasi-n00b LM1875 problems
« Reply #18 on: October 06, 2013, 10:52:41 PM »
Yes, that drawing showing the 100uF's dotted is misleading and confusing.

Where the filter caps for the supply are fitted very close to the chip amp then no additional bypassing may be required.  In the case of a vehicle supply(*), or when a mains supply may be situated more than a few inches away, then local bypass caps, both high frequency 0.1uF and low frequency 100uF are required near the chip, with a minimal loop length through the chip and bypass caps.

The whole object of bypass caps is to hold the supply rails at AC ground by providing a low impedance path from the chip supply pin to the chip ground pin (or AC grounded pins).  Given the performance of some of these chip amps "low impedance" can turn out to be quite critical and demand that the bypass caps be fitted right next to the chip and connected with really fat PCB traces or even heavy wire.

(* While vehicle supplies normally have a very low source resistance for DC, allowing very high currents to flow, their AC impedance often leaves a lot to be desired in amplifier terms, hence those very large "filter" caps that get sold, which are actually bypass caps and should be fitted right at the amp, not in the engine bay.)

Bypassing is bypassing, so in the case of your LFO it will also be ensuring that the supply rails are effectively shorted to ground for AC ('tho a LFO will normally require rather larger value caps to fulfil this requirement).

The background reason for this need for bypassing is that if a supply rail isn't properly bypassed, shorted to ground for AC by a capacitor, then AC signals will be able to pass along the supply from the output stages to the input stages and create instability, positive feedback (or possibly a reduction in gain due to negative feedback, but instability is more common).

This is still true even where the input and output circuits are within the same case, as with many chip amps.

If you have high and low frequency bypasses fitted right at the chip you can be a whole lot more relaxed about feeding power to the chip amp from a power supply some distance away.
If you say theory and practice don't agree you haven't applied enough theory.