The (high-cap.) unregulated PSU for chipamps

Status
Not open for further replies.
Jump to Latest
#561
Mick_F said:


I dont agree. Indeed additional L wont make a problem for the line resistors R3 and R4. For the snubber resistors R5 and R6, on the other hand, I think it is more critical. The network has proven to be very sensitive to small variations in R and C and hence may be also very sensitive concerning an extra L. The values used for R and C are very small, and hence the addition of some L may lead to a small RLC serial oscillator.

Click to expand...

Indeed. The resistors I was referring to were those between the caps on the raw supply. Adding an L to a zobel is moving the affected bandwidth.


Carlos
 
#563
R 1-4

Thank-you everybody for your assistance. This is what I "sort of" thought, but my electronics skills are basic.

Carlos FM has also written two replies to me which he said I can post for general edification:

Post 1 ...
I don't know if you missed it, but I updated the PSU. R3 and R4 are now 1R/10W wirewound resistors. Inductance in series is fine, what I would avoid is inductive resistors to ground. So, R1 and R2 must be 2 or 3W carbon resistors.

Post 2 ...
[I would] just like to add that (as I said several times) the snubber resistors must NOT be wire wound types too, and the snubber cap must be a small one. No big/high voltage polyprops here high inductance!), 100V rating MKT is what I recommend.
(End of Quote from CarlosFM)

This is in accordance with what you have advised me, but highlights that R1 and R2 (across the 4700uF caps to ground) must be low inductance types.

Hope this helps ... it did help me.

Regards,
George.
 
#566
Mick_F said:


I dont agree. Indeed additional L wont make a problem for the line resistors R3 and R4. For the snubber resistors R5 and R6, on the other hand, I think it is more critical. The network has proven to be very sensitive to small variations in R and C and hence may be also very sensitive concerning an extra L. The values used for R and C are very small, and hence the addition of some L may lead to a small RLC serial oscillator.

Mick
Click to expand...
You have a "smidgen" in the power supply regardless -- it's the leakage inductance of the transformer secondary -- and it is at least 2 orders of magnitude greater than the ESL of any decent electrolytic.

trying to "effect" the realizable impedance of a power supply by focusing on the trivial effect of filter capacitor ESL -- you would think that King Canute finally rolled back the waves of physics.

here's the real power supply which you can analyze with virtual instruments -- in LTSpice.

An externally hosted image should be here but it was not working when we last tested it.
 
#567
Ok, I let my ears decide.

Yesterday, I have replaced the resistors on the positive rail side by wirewounds and on the negative side I kept the carbons. The first listening impression was a shock. On the negative rail side, there was a lot of increased bass authority with more punch, a much wider soundstage and transparent clarity in the treble.

Mick
 
#569
I have a question that I need to ask, just to be sure 🙂
It's about the text in the lower part of the schem.
the 100nF goes parallell to the 2200uF capacitor and the 330nF goes between V+ and V- (that's pin 1 and 4, right?)

and the snubber concisting of the 0,47R in series with 47nF also goes between pin 1 and 4, because "Just before these caps" refers to the 330nF or am I wrong?

best regards

EDIT:
I just saw that there are no 2W 0.47R carbonresitors available, nor is there any metalfilm. The only alternative is larger wirewound resistors.

So what is the recommendation, either skip the snubbernetvork on he pins or to go with the wirewound?
 
#571
kmj, I don't think you are looking at the latest versions of the PSU and amp.

Here's the files, I hope it's clear(er) now.
Click to expand...

Thanks, as you thought, wrong version. This made everything clear.
However, it still doesn't solve my other problem thou. I can't get ahold of those kind of resistors, nothing below 0.56R 1W carbon (composition?) or 1R metalfilm.

Guess that i could spend some time asking around different forums and hope that someone has some 4sale 🙁
 
#574
kmj, one way to find out how high power rating these resistors must have is to estimate the ripple and then make a simple calculation. Soon you will discover that the power dissipated in these resistor are uW! => almost any holemounted resistor will do.

1R//1R metallfim 1% 0.6 W will probably do fine.

EDIT: Seen that Mick thinks the same 🙂
 
#575
For the 0R1 resistors on the amp snubbers I made some resistors by myself. I will make some description and pictures of the procedure later and post them here.....
Click to expand...

Now that sounds interesting 😀
It's those thats the hardest to find, those for the psu can probably be substituted by putting some resistors in parallell as you said.
 
#577
peranders

I missed your reply but thanks!
I acually thought of resistors in parallell but it was the 0.1R that messed things up but I found some 0.22R powerresistors that should do, I just wonder what kind or resistors i'll be getting 🙂, they didn't specify that.

Thanks again guys.
 
#579
Well, I don't think that I'll take you up on that bet 🙂
Well, i'll manage somehow, but I REALLY hate to have to put stuff in parallell, that's why i'm trying to worm my way out of it :angel: . Don't ask why but somehow it's really annoying.......
 
#580
kmj, sorry for the delay, too much work...

My schematics have 1W resistors for the snubbers because that's what I can find more easily at low values.
I 'standardized' on this because I think that if you search for 1W resistors below 1 ohm you are more likely to find them.
I have 0.47R resistors at 1/2W (carbon, 5%) and also 0.6W film (1%).
Below 0.47R I can only find (locally) 1W.
1W 5% carbon I have 0.47R, 0.22R and 0.1R.
I also have 3W carbon resistors on all these values, but they are big.
 
Status
Not open for further replies.
Top Bottom