Carlos' snubberized Gainclone Power supply, Part II

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
As we all know BrianGT is now marketing Carlos snubberized Gainclone PSU
http://www.chipamp.com/images/ps-rev3-sch.gif

I started the snubberization thread here . The question them is the same as now, almost anyway. I thought it was a bit peculiar to have 100 nF in parallel with 100 nF + 1 ohms. We have seen from real measurements and simulations that this unsnubbed 100 nF isn't doing any good (my words). This single 100 nF creates a peak, not big but still a peak... so what does the gathered expertise say about my "enhanced power supply"?

What does Carlos and Joseph_K say about this, I mean in theory.
 

Attachments

  • snubberized_psu.pdf
    24.9 KB · Views: 3,567
P-A, my psu is made like that, except that my snubb resistors are 2ohm. I did it like that because I wanted to keep the capacitors kind of separated for each chanel.
I dunno if it sounds better or not. OK, to me it does not sound worse in any way than the basic Carlos configuration, I could well be wrong though.

I am keen on seeing some theory on how the two configs differ from eachother :)
 
jaudio:
Nope. I am using only the snubbers, the 100nf bypass caps did nothing to the sound ( but they did something negative to the looks of the psu) with the 4 BC components 3300uf filter caps I am using per channel right now.

I like big differences, I tend to get frustrated with myself if I try to hear less than obvious differences.

Or did you mean if I use the same type of 100nf caps? I use mkt.
 
No, I didn´t.
If I would stay true to Carlos values I guess that I would have to use 50nf. I should perhaps try it.

Edit: I also need to be clear. I did what I did because I wanted to use a single bridge (consisting of 8 diodes) but have sort of separate filter capacitors for each channel, so I felt that I would need separate snubbers for each channel, even though they are not isolated from each other. I just ran double wires after the bridge ( taking guesses at how electricity flows and so on).

Perhaps this is madness, thats why I am keen of hearing some theory from those who have the knowledge.
 
postpunk said:
jaudio:
4 BC components 3300uf filter caps I am using per channel

That's a coincedence, 2x3,300uF BC components per rail (4 per channel) is exactly what I use! And it sounds really nice. I have parts to try this 'snubber', but for some days I've been trying to learn the sound of the amp as is.

I started with just 1x3,300uF per rail, and going to 2x3,300uF has led to a more relaxed, natural sound with easier detail, better low bass, and more realistic dynamics. More musical. The downside (apart from cost) is a reduction in the freshness and vitality of sound. I wonder if the 'snubber' somehow affects this quality. Bypass at the chip has always been 100uF Rubycon ZA. I will share my listening impressions after trying the snubber, and maybe some other ideas.
 
peranders,

You are clearly committed to improving your understanding of the power supply in an amplifier, and how to improve it. You have mentioned very little of your own listening impressions. Do you tend to hear differences when you change the filter capacitence, any bypass caps, bridges, transformers etc.?

If you do not, then perhaps you will never hear the difference adding this resistor and capacitor might make.
 
Sorry, can't give you any information about audible results or conclusions.

But if we speak technically I think the C2, C7 (in my schematic in the first post) must be eliminated becuase when you have them there unsnubbed you will get the "horrible" reasonance peak.

The purpose of inserting a 1 ohms resistor is to eliminate this peak.
 
I tried it last night.... and the initial feeling is that it works... and very well! I don't think I implemented it especially well, but clearly well enough to make a change for the better.

I used 1R/3w (ww) with 82nF/63v (polyester). I used no film cap straight across the rails. Still have 100uF Rubycon as bypass at chip's pins. Rails are 2x3,300uF/40v BC Comp. All wired p-p, 2 large packaged bridges per channel, twin secondary 100va 25v trafo/channel. Running off 3kva isolation transformer. Listening to: Joni Mitchell - Hits, Young Gifted & Black (various), and Neil Young - Harvest.

Initial listening (I'll only trust this after several days) showed a marked decrease in fuzz and haze, allowing ambient acoustic cues and fine detail to be much more clearly audible. Voices are more intimate, the bass is more tuneful, the top end is less messy. The sound is not the same as using just one 3,300uF cap per rail- this sound is more realistic and more sophisticated. It sounds sweet, controlled, and 'together'.
 
I did a simple simulation and I got this:

50 V peak AC, 20kHz (in) 10 ohms load, 10000 uF + 100 nF+ 1 ohms as snubber. Then I got 45 mV peak over the 1 ohms. => 2 mW

It's possible that 50 Hz in and 20 kHz half pulses (class B) creates different results but I'm a newbeginner when it cvomes to simulation.

... so I'll doubt that you will need several watts in the resistors.
 
Upupa Epops said:
To Simont : Why are you using 3 W resistor ? It's technical nonsense, 'cos there you need only microwatt dissipation. And whatabout to try connect bypass condenser between rails ? Disturbing on rails have opposite phase, so by this cap will be shorted. :cool:

peranders said:
I did a simple simulation and I got...2 mW

Ok, thanks for bringing this up. I'll just use the usual 1/4w carbon next time... and if I get :hot: then I'll know who to blame :clown: I'm not a techincal type, I just follow like the other sheep ;) "baaah"

Upupa - are you saying I might benefit from the bypass film cap? What should I try? I have lots of 330nF and 82nF polyester to hand. Also have some Panasonic FC 33uF, but I guess their parameters could be swamped by the other caps in the circuit :confused:
 
SimontY said:




Ok, thanks for bringing this up. I'll just use the usual 1/4w carbon next time... and if I get :hot: then I'll know who to blame :clown: I'm not a techincal type, I just follow like the other sheep ;) "baaah"

Upupa - are you saying I might benefit from the bypass film cap? What should I try? I have lots of 330nF and 82nF polyester to hand. Also have some Panasonic FC 33uF, but I guess their parameters could be swamped by the other caps in the circuit :confused:


if the amplifier clips, there is a lot more power (than a microwatt) being dissipated by the resistor !!!!!

other thing to note -- if there is an overvoltage situation (which will happen with an inductive load like a speaker or motor -- the reverse voltage can substantially exceed the rail voltage -- the inductor acts like the storage inductor in a switching power supply -- National discussed this in an application note in 1993 -- the chip itself is protected from the over-voltage -- the snubber isn't.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.