I'm planning on using the speaker-turn-delay-dc-protector-boards with the multi channel Leach amp I'm building. In his documentation of his design, Mr. Leach explains:
"R49, L1, R50, and C25 suppress parasitic oscillations that could be induced by shunt capacitance in the loudspeaker load. R50 and C25 mount on the loudspeaker output binding posts. On the first amplifier I built, I had these on the circuit board where they caused violent oscillations because the current through R50 and C25 generated positive feedback when it flowed through the circuit board signal ground lead. When the oscillations occurred, the heat sinks would get very hot. Moving R50 and C25 to the loudspeaker output binding posts solved this problem." R50 = 10 ohm, 2 W, C25 = 0.1 uF, 100 V film.
(The Leach Amp - Output Stage)
I'm concerned about the possibility of inducing oscillations with adding the speaker protector boards - is anyone here able to comment on their likely effect?
"R49, L1, R50, and C25 suppress parasitic oscillations that could be induced by shunt capacitance in the loudspeaker load. R50 and C25 mount on the loudspeaker output binding posts. On the first amplifier I built, I had these on the circuit board where they caused violent oscillations because the current through R50 and C25 generated positive feedback when it flowed through the circuit board signal ground lead. When the oscillations occurred, the heat sinks would get very hot. Moving R50 and C25 to the loudspeaker output binding posts solved this problem." R50 = 10 ohm, 2 W, C25 = 0.1 uF, 100 V film.
(The Leach Amp - Output Stage)
I'm concerned about the possibility of inducing oscillations with adding the speaker protector boards - is anyone here able to comment on their likely effect?
adding zobel networks to leach amp w speaker protection boards
Can I impose on you to share more about how that will help? I'm guessing it's based on experience, but knowing would add a confidence factor. Isn't R50 & C25 a zobel network - does it matter if it's too the left or right of L1? I'm worried that adding one to the left of L1 would induce the same condition Mr Leach mentioned with oscillations on his first revision of the amp.
Mr Leach also wrote - "I have seen amplifier circuit diagrams with different variations of the circuit formed by R49, L1, R50, and C25. Some have R50 and C25 on the circuit side of R49 and L1. Some have a series resistor and capacitor to ground on both sides of R49 and L1."
He doesn't explain his reasons for his final design beyond this - are you able to explain it? And how this applies to the addition of the circuit protection boards?
(Looking at your other posts here I can easily see where you have a wealth of experience - please don't think my questions are doubting your skills, it's just that the whole point for me of building my own amps was in learning something.)
Thanks!
Can I impose on you to share more about how that will help? I'm guessing it's based on experience, but knowing would add a confidence factor. Isn't R50 & C25 a zobel network - does it matter if it's too the left or right of L1? I'm worried that adding one to the left of L1 would induce the same condition Mr Leach mentioned with oscillations on his first revision of the amp.
Mr Leach also wrote - "I have seen amplifier circuit diagrams with different variations of the circuit formed by R49, L1, R50, and C25. Some have R50 and C25 on the circuit side of R49 and L1. Some have a series resistor and capacitor to ground on both sides of R49 and L1."
He doesn't explain his reasons for his final design beyond this - are you able to explain it? And how this applies to the addition of the circuit protection boards?
(Looking at your other posts here I can easily see where you have a wealth of experience - please don't think my questions are doubting your skills, it's just that the whole point for me of building my own amps was in learning something.)
Thanks!
Last edited:
I'm not an expert to explain you why a zobel is needed and where it has to be inserted. I use a so called "pi" network at the output stage as described by at least Bob Cordell.
In my thread there are many small discussions why and where to put a zobel - simply search the thread for "zobel":
http://www.diyaudio.com/forums/soli...ormance-class-ab-power-amp-200w8r-400w4r.html
What I know from simulation: if the zobel isn't placed directly at output stage the PM and GM of the amplifiers worsens hence the stability of the amplifier betters if you add the zobel on the left side of the output inductivity.
BR, Toni
In my thread there are many small discussions why and where to put a zobel - simply search the thread for "zobel":
http://www.diyaudio.com/forums/soli...ormance-class-ab-power-amp-200w8r-400w4r.html
What I know from simulation: if the zobel isn't placed directly at output stage the PM and GM of the amplifiers worsens hence the stability of the amplifier betters if you add the zobel on the left side of the output inductivity.
BR, Toni
Thanks for such helpful information - I'd already looked a bit at your 2stageEF thread, I came across it when I looked at your posts here (kind of hard to miss at 1800+ posts on just one thread!). There's so much good information in it I'll be mining it for a long time, not just for information on zobel networks. I really need to learn what sort of conditions I'll need to use to test the stability, I suspect given the level of detail I've seen in other parts of that thread that I'll find that information there too.
I'm attempting to achieve similar levels of power density to what you're doing (130W8R, but 5 channels). While I have the PCBs from the DIY store for the PS and speaker protection, your housekeeping circuits have some nice features like auto power off I like - I'll send you an email to the address in that thread about availability of them.
I'm attempting to achieve similar levels of power density to what you're doing (130W8R, but 5 channels). While I have the PCBs from the DIY store for the PS and speaker protection, your housekeeping circuits have some nice features like auto power off I like - I'll send you an email to the address in that thread about availability of them.
Neville Thiele described two Networks that he suggested be added to an amplifier output to help with stability and obtaining adequate margins.
a.) The first version which is by far the most popular shows a resistance in series with a capacitor (R+C) from output Hot to Main Audio Ground. This is followed by an inductor with a parallel resistance (L||R)
b.) the second version starts with the L||R in the output hot lead and then has a capacitor from the output hot to main audio ground. This is similar to the version adopted by Prof Leach. It is rarely implemented.
You can apply a very simplified analysis of those two alternative Thiele Networks by lookng at the DC effect and the UHF effect of those components to the signal that needs to pass.
For a.
@ DC the capacitors are effectively open circuit. The Thiele Network a. simplifies to an inductor with a tiny resistance in series with the load. i.e. 99.99% of the DC gets through to the load.
@ UHF the capacitor simplyfies to a short circuit and the inductor is effectively an open circuit.
The R+C becomes an R at UHF and the L||R becomes a tiny R at UHF.
The amp output sees an R to Main audio ground and a further R in series with any attached load. That attached load will be an effective open circuit at UHF so the only load the amplifier sees is the R of the R+C. The cables could provide a bit of additional C that brings this last R back into the current route and would be in parallel to the first R. This route for output current ONLY applies at frequencies above the audio band and at UHF demands that all inductance in the route be kept very low. Otherwise the route becomes ineffective in the band where it should pass current.
For b.
@ DC the L||R becomes a very low resistance. The C across the output becomes an open circuit. All we have is a small R in series with the load and 99.99% of the DC gets through to the load. (same as version a.)
@ UHF the inductor becomes an open circuit in parallel with it's R and the C across the output terminals becomes a short circuit. The amplifier output sees R of the L||R as the load into the shorted output terminals. Again this is the same loading as was provided by the a. version.
At the two extremes of frequencies i.e. at DC and @ UHF a. and b. simplify to become the same loading for the amplifier output.
Dr Cherry discussed this in a paper published in 1995.
He pointed out that the two Thiele Networks were the limits of a continuum that had an infinite variety of versions.
I posted a spreadsheet years ago that used the Cherry formulae with a reference to the Cherry article and two years later (1997) a short update.
Here it is again.
Try to get hold of the Cherry paper.
The sole purpose of either version of the Thiele Network is to provide a defined load for the amplifier at frequencies above the audio band. Any added inductance in those routes to the Capacitors and back to the source of those HF currents will impede the effectiveness of the Thiele Network. Version b. must of necessity have longer routes and thus have higher inductance. This may make version b. less effective at the extremes of HF. That may be a reason why most Designers avoid this version.
Finally:
I have suggested a few times in the past (predating R.Cordell) that we combine the a. and b. versions to create a Pi filter.
The main purpose of the b. part is to attenuate interference that is injected by the speaker cable.
Dr. Cherry in a separate paper suggests moving the tapping point of the NFB from the output of the amplifier to after the R of the R+C. This adds a further filter of interference before the NFB signal is sent back to the -IN pin of the amplifier. I showed this Cherry alternative tapping in a Thread many years ago. Few (any?) have made any comment on the Cherry idea for the alternative tapping.
a.) The first version which is by far the most popular shows a resistance in series with a capacitor (R+C) from output Hot to Main Audio Ground. This is followed by an inductor with a parallel resistance (L||R)
b.) the second version starts with the L||R in the output hot lead and then has a capacitor from the output hot to main audio ground. This is similar to the version adopted by Prof Leach. It is rarely implemented.
You can apply a very simplified analysis of those two alternative Thiele Networks by lookng at the DC effect and the UHF effect of those components to the signal that needs to pass.
For a.
@ DC the capacitors are effectively open circuit. The Thiele Network a. simplifies to an inductor with a tiny resistance in series with the load. i.e. 99.99% of the DC gets through to the load.
@ UHF the capacitor simplyfies to a short circuit and the inductor is effectively an open circuit.
The R+C becomes an R at UHF and the L||R becomes a tiny R at UHF.
The amp output sees an R to Main audio ground and a further R in series with any attached load. That attached load will be an effective open circuit at UHF so the only load the amplifier sees is the R of the R+C. The cables could provide a bit of additional C that brings this last R back into the current route and would be in parallel to the first R. This route for output current ONLY applies at frequencies above the audio band and at UHF demands that all inductance in the route be kept very low. Otherwise the route becomes ineffective in the band where it should pass current.
For b.
@ DC the L||R becomes a very low resistance. The C across the output becomes an open circuit. All we have is a small R in series with the load and 99.99% of the DC gets through to the load. (same as version a.)
@ UHF the inductor becomes an open circuit in parallel with it's R and the C across the output terminals becomes a short circuit. The amplifier output sees R of the L||R as the load into the shorted output terminals. Again this is the same loading as was provided by the a. version.
At the two extremes of frequencies i.e. at DC and @ UHF a. and b. simplify to become the same loading for the amplifier output.
Dr Cherry discussed this in a paper published in 1995.
He pointed out that the two Thiele Networks were the limits of a continuum that had an infinite variety of versions.
I posted a spreadsheet years ago that used the Cherry formulae with a reference to the Cherry article and two years later (1997) a short update.
Here it is again.
Try to get hold of the Cherry paper.
The sole purpose of either version of the Thiele Network is to provide a defined load for the amplifier at frequencies above the audio band. Any added inductance in those routes to the Capacitors and back to the source of those HF currents will impede the effectiveness of the Thiele Network. Version b. must of necessity have longer routes and thus have higher inductance. This may make version b. less effective at the extremes of HF. That may be a reason why most Designers avoid this version.
Finally:
I have suggested a few times in the past (predating R.Cordell) that we combine the a. and b. versions to create a Pi filter.
The main purpose of the b. part is to attenuate interference that is injected by the speaker cable.
Dr. Cherry in a separate paper suggests moving the tapping point of the NFB from the output of the amplifier to after the R of the R+C. This adds a further filter of interference before the NFB signal is sent back to the -IN pin of the amplifier. I showed this Cherry alternative tapping in a Thread many years ago. Few (any?) have made any comment on the Cherry idea for the alternative tapping.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.