Help!! Irritating sound of my GainKlone

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Nuuk said:
When building prototypes, don't just leave everything just sitting there. Screw the transformer onto a chunky pice of timber, the rectifier bridge too and even fix some of the cables down.:nod:


Yes - after some recent mods I was left with a bit of a lash up. odd bits of cheap wire here & there... I thought it was sounding good until I did it all properly this w/e with careful layout, earthing etc with cat 5. The improvement was suprising ( noticed immediately by my other 1/2 ! )

cheers

mike
 
weak transformer

For me, athough it may not be the only responsible for the sound you have, your transformer is very weak.
You could get another one and use two, one for each channel.

Hey Kwei, nice to see you here.:cool:
I wonder what you use as a power amp...:scratch:
If it's the same league as your preamp, it has to be something special.:nod:
 
Re: Re: Help!! Irritating sound of my GainKlone

Kuei Yang Wang said:


Finally, the two parallel capacitors before and after the bridge rectifier are SERIOUSLY bad news in my view. remove the 0.47uF Cap and place it after the Bridge to not loose it's use.

Sayoanra

For many years I put a cap like this ( often upto 4.7u !!! ) in this position and without really knowing what it was doing I noticed a HUGE improvement in the sound. ( Also the transformer warmed up a bit !!! )

Having done some spice research on this recently, I now realise that this cap has the effect of drastically reducing the frequency, but not the amplitute and lenght, of the ringing in low resistance secondaries.

My current idea is to keep the 4.7uF with 100ohm in series and .47uF across them both. This stops the ringing and removes and fast transients that might create any RF nasties.

I hope this does not sound too wacky:bigeyes:

If it does here is the test cct I used - try it and see what you think

cheers

mike

PS the voltage source is 18V AC
 

Attachments

  • capture2.jpg
    capture2.jpg
    14 KB · Views: 503
tiroth said:
I think you are hard at work on reproducing Calculating Optimum Snubbers.

Yes - this was my starting point, but as you can see it was not my finnishing point;)

based on my previous experience I think the bigger value caps will sound better. If the 1st cap is about 10 times smaller than the second its is quite easy to find a resistor value that will stop the ringing.

with 10uF & 100uF+10ohms the tendency for ringing simply merges into the next mains cycle - this is probably going a bit too far but who knows it may sound best of all as it is also quite a powerful mains noise filter !

cheers

mike

to see what is going on put a current meter in between the snubber cct and the TR coil and a voltmeter between ground & snubber caps.
 
Re: weak transformer

Koinichiwa,

carlosfm said:

I wonder what you use as a power amp...:scratch:
If it's the same league as your preamp, it has to be something special.:nod:

You can see the whole system here:

http://www.arduman.com/aa/Sayfalar/thorsten/thorsten.htm

I do think the Poweramps are in the same league as my Pre. But that is just my opinion. A look inside:

amp_bott.jpg


A look from above:

amps.jpg


And a schematic for those who like that sort of thing:

amp_schem.gif


Sayonara
 
Let’s see what I am going to do

1. Remove the wima bypassing caps to the 1000uF electrolytic
2. Move the RifaX2 caps after the bridge rectifier
3. Add 0.22ohm white coffin wire wound resistors between the transformer and the bridge rectifier. Insert two Wima 10nF cap before and after the wire wound resistors.
4. Improve the wiring + arrangement (nail them on a board), transformer upgrade at later stage.
5. CONVERT to inverting configuration. HALILUYA!!!
6. Lowering the value of the resistors used in the amp circuit.
7. By pass the input 4.7uF cap.
8. Add 0.22 ohm Power film after the out pin and the binding post. I think I need a heat sink for this resistor or I have to attach it to the chassis of the amp. So far I can only find Meggritt MPR20 with this value in my country.
9. Upgrade Transformer to 400 or 500VA, +/-24Vac

I will also try the sunbber circuit suggested by mikelm.

I start off with the most cost effective change.

So far I had removed the bypassing wima caps and one set of RifaX2+FKP1 combo before the bridge rectifier, the other combo is still in the circuit. The effect: the bass region of the sound clear up quite bit. Previously, I perceived the bass to be good, but a little on the face, too up front. Now the bass have cleaned up, the on the face feeling is gone, the music is more relax and at ease, but still not quite there yet. Another thing I have noticed, with the bass cleaned up, the mid and the high are also clearer, the high is also more extended and slightly more prominent. I would say the sound before the removal of the caps is more muffle. The improvement is positive.

Kuei’s diagnosis is spot on. Alikato! Kuei San.

Again thank you everyone. I will feedback the effect of each change once I have done them.


Regards,
SW Yeo
 
Let’s see what I am going to do

>3. Add 0.22ohm white coffin wire wound resistors between the transformer and the bridge rectifier. Insert two Wima 10nF cap before and after the wire wound resistors.

>5. CONVERT to inverting configuration. HALILUYA!!!
>6. Lowering the value of the resistors used in the amp circuit.

If you change to inverting mode the input resistor IS the i/p resistance so if you lower your values you will have to be sure that our preamp can drive this value. For inverting mode many of us get great results with 10K/220K

>8. Add 0.22 ohm Power film after the out pin and the binding post. I think I need a heat sink for this resistor or I have to attach it to the chassis of the amp. So far I can only find Meggritt MPR20 with this value in my country.

should be ok without the sink - but might sound good on it

>I will also try the snubber circuit suggested by mikelm.

you will need this OR coffins & caps ( point 3 ), not both !

Sounds like you will end up will a great sounding amp - good luck

mike
:)
 
Re: Re: weak transformer

Kuei Yang Wang said:
Koinichiwa,

You can see the whole system here:

http://www.arduman.com/aa/Sayfalar/thorsten/thorsten.htm

I do think the Poweramps are in the same league as my Pre. But that is just my opinion. A look inside:

amp_bott.jpg


A look from above:

amps.jpg


And a schematic for those who like that sort of thing:

amp_schem.gif


Sayonara


This guy is fantastic.:nod:
Kuei, I'm making an ultra-high-end:eek: :eek: preamp.
Not in the same league as yours, that's outer-space high-end.:bigeyes:
Reports soon.
 
Yesterday, I replaced my 220 k feedback resistor with 330 k. Also I use that value on the + input. (DC <0.7 mV !)

I had quite some stability issues recently after some changes. But the amp seems to be fully stable now again and also sounds better. More dynamic, better defined and more gain. Everything better :)

Fedde
 
Koinichiwa,

overmind said:

1. Remove the wima bypassing caps to the 1000uF electrolytic

Yes. Bypassing is a dicey game. I know people who routinely use just a single bypass. If your "lwo-Z" Electrolytic cap has a "normal" parasitic inductance (say 300nH - gives 180mOhm @ 100KHz and you a 470nF bypass you get a high Q resonant circuit at around 420KHz, quite possibly in a region where it can add to the instability o the circuit.

overmind said:

2. Move the RifaX2 caps after the bridge rectifier
3. Add 0.22ohm white coffin wire wound resistors between the transformer and the bridge rectifier. Insert two Wima 10nF cap before and after the wire wound resistors.

Yup. BTW, this basic suggested snubber can be much improved upon, I suggest the late John "Buddah" Camilles article in Sound Practices on "Reducing Diode Noise" for a really comprehensive treatment how to kill the various rectificaion noise sources....

overmind said:

4. Improve the wiring + arrangement (nail them on a board), transformer upgrade at later stage.
5. CONVERT to inverting configuration. HALILUYA!!!
6. Lowering the value of the resistors used in the amp circuit.

Yup.

overmind said:

7. By pass the input 4.7uF cap.

As before, beware of "Bypassing". It is a good tool to be applied correctly, but failure to do so can often make things worse.

overmind said:

8. Add 0.22 ohm Power film after the out pin and the binding post. I think I need a heat sink for this resistor or I have to attach it to the chassis of the amp. So far I can only find Meggritt MPR20 with this value in my country.

Meggit resistors and chassis will be fine.

overmind said:

9. Upgrade Transformer to 400 or 500VA, +/-24Vac

I would strongly suggest to go for four separate lower power transformers. That way you can operate them in the "low flux" connection (1/2 rated primary voltage), plus I'd suggest adding the DC killer circuit. Using fully separate supplies per channel comes highly recommended too. Using the right mechanical arrangement for the transformers can in addition minimise any leakage from the mains into the actual circuit, by the expedient of in effect "bucking out" the leakage components.

overmind said:

I will also try the sunbber circuit suggested by mikelm.

The type of snubber Mike suggests as well as the even more extensive ones by John Camille need to be accuratly calculated for the given mains transformer - they are NOT universal. Getting them wrong is worse than not having them.

The simple CRC PI Filter recommended above is nowhere near as effective, yet it is much less critical and close to "universal". You could try using a large value capacitor on the "rectifier" side of this CRC circuit, have a listen.

Sayonara
 
where can I find this article

Kuei Yang Wang said:

I suggest the late John "Buddah" Camilles article in Sound Practices on "Reducing Diode Noise" for a really comprehensive treatment how to kill the various rectificaion noise sources....

I would like to read this. I could not find it on the sound practices site / back issues.

do you or anyone know where I can find it ?

Mike

p.s. Now, if Kuei Yang Wang is chinese, Kuei would be your family name and Yang Wang your given names ( or something similar. ) So should I be calling you 'Mr Kuei or Yang Wang or just Wang ? Hope you don't mind me asking ;)
 
Re: where can I find this article

Koinichiwa,

mikelm said:


I would like to read this. I could not find it on the sound practices site / back issues.

do you or anyone know where I can find it ?

It's in issue 7. But do buy the CD-Rom - it is worth having, printing out and reading....

mikelm said:


p.s. Now, if Kuei Yang Wang is chinese, Kuei would be your family name and Yang Wang your given names ( or something similar. ) So should I be calling you 'Mr Kuei or Yang Wang or just Wang ? Hope you don't mind me asking ;)


Non of the above. Please see:

http://www.sazan.net/digest/sazguide.shtml#Kaiyanwang

Kuei Yang Wang is another possible rendering of the name of the "Demon-Eye King".

Sayonara
 
Kuei Yang Wang said:
Koinichiwa,

Yes. Bypassing is a dicey game. I know people who routinely use just a single bypass. If your "lwo-Z" Electrolytic cap has a "normal" parasitic inductance (say 300nH - gives 180mOhm @ 100KHz and you a 470nF bypass you get a high Q resonant circuit at around 420KHz, quite possibly in a region where it can add to the instability o the circuit.


Yup. BTW, this basic suggested snubber can be much improved upon, I suggest the late John "Buddah" Camilles article in Sound Practices on "Reducing Diode Noise" for a really comprehensive treatment how to kill the various rectificaion noise sources....


I think I have a lot to read up. Thanks for pointing to the link/source.



Meggit resistors and chassis will be fine.

I bought some 1 ohm carbon film resistors, 1 watt rating, together with the white coffin resistors. I intend to parallel 5 pieces of the 1 ohm resistors to achieve 0.2ohm.

If the purpose of using Film Power resistor is for it's non inductiveness, then would carbon film fit in as well? Paralleling them would further reduce the overall inductance.


I would strongly suggest to go for four separate lower power transformers. That way you can operate them in the "low flux" connection (1/2 rated primary voltage), plus I'd suggest adding the DC killer circuit. Using fully separate supplies per channel comes highly recommended too. Using the right mechanical arrangement for the transformers can in addition minimise any leakage from the mains into the actual circuit, by the expedient of in effect "bucking out" the leakage components.

Four transformers would much costly than one large transformer. I have the schematic of your inverted Gainclone, the verison with 4 transformers for one channel.

Could I compromise by having all four set windings (4 x primary:230V + secondary:24V) into a transformer or I have to keep them as four individual transformers?



The type of snubber Mike suggests as well as the even more extensive ones by John Camille need to be accuratly calculated for the given mains transformer - they are NOT universal. Getting them wrong is worse than not having them.

The simple CRC PI Filter recommended above is nowhere near as effective, yet it is much less critical and close to "universal". You could try using a large value capacitor on the "rectifier" side of this CRC circuit, have a listen.

Sayonara


I have inplemented the CRC PI filter. With the 10watt 0.22ohm white coffin + 10nF 1.6kV FKP1. The effect is positive, thus music becomes more relaxing, less harsh and more enjoyable.

It is possible to go for CRCRC, 10 watt, 0.1 ohm +10nF?


Thank You
Yeo
 
Koinichiwa,

overmind said:

I bought some 1 ohm carbon film resistors, 1 watt rating, together with the white coffin resistors. I intend to parallel 5 pieces of the 1 ohm resistors to achieve 0.2ohm.

If the purpose of using Film Power resistor is for it's non inductiveness, then would carbon film fit in as well? Paralleling them would further reduce the overall inductance.

The purpose is actually for non-magnetic resistors. The "Powerfilm" types I have tried where all non-magnetic.

overmind said:

Four transformers would much costly than one large transformer.

True, but not that much. If you start to go looking for Surplus Transformers you might find something easily.

overmind said:

I have the schematic of your inverted Gainclone, the verison with 4 transformers for one channel.

Two transformers for one channel, actually.

overmind said:

Could I compromise by having all four set windings (4 x primary:230V + secondary:24V) into a transformer or I have to keep them as four individual transformers?

The fundamental idea of having the windings in series is to halve the peak flux in the Transformer core. This helps making the transformer quieter and allows in the better power transfer despite the doubled DCR.

If you have Transformers custom wound you can have them made with more primary turns (at least 30% more)to reduce the flux and have a small airgap designed into the transformer, this would achieve pretty much the same.

If you get something custom made it's also worth asking for fully symmetrical primary and secondary windings (which suggests a C-Core or R-Core type transformer) and electrostatic screens.

overmind said:

I have inplemented the CRC PI filter. With the 10watt 0.22ohm white coffin + 10nF 1.6kV FKP1. The effect is positive, thus music becomes more relaxing, less harsh and more enjoyable.

It is possible to go for CRCRC, 10 watt, 0.1 ohm +10nF?

Sure. You can do that, or you could place a second CRC after the rectifier. Also, by all means experiemnt with the value of the second capacitor in each CRC Filter (looking from the transforer to the output).

Sayonara
 
Hi Quei

I am sure you have explained this before but wouldn't just two transformers (not centre-tapped), primaries and secondaries wired in series not achieve the low core flux? What is the advantage of having 4 separate windings? Reduction of external flux? Do you have to observe particular relative positioning of the transformers? Would that work with EI cores? Just interesting if you have actually listened to these effects and would care to share.
And last question. Is it possible to use split bobins (on EI)? Would the reduction in capacitance between primary and secondary be beneficial?


thanks

peter
 
Koinichiwa,

analog_sa said:

I am sure you have explained this before but wouldn't just two transformers (not centre-tapped), primaries and secondaries wired in series not achieve the low core flux?

Sure.

analog_sa said:

What is the advantage of having 4 separate windings?

The ability to use standard off the shelf Mains transformers? All the better Torrids I know come from the factury with 115V+115V primaries and xxV+xxV secondaries. That is just how they are being made.

analog_sa said:

Do you have to observe particular relative positioning of the transformers?

If you wish to use two seperate transformers with opposite phase connections of the primaries to cancel the "leakage" into chassis and secondaries (as I suggest), you must ensure the transformers are arranged as symmetrical as possible.

I also suggest using non-metallic fixings if you use metallic chassis. You can simply place the transformer o a suitably thick piece of rubber, place a large nylon screw (or several for heavy transformers) well but thinly covered with that special seperation grease and then poor a suitable acrylic resin or the like (fairly thick consistency) into the transformer center. The result - a transformer center solidly gooped together with screw theread(s) for your nylon screw(s).

analog_sa said:

Would that work with EI cores?

Sure. Both low flux and killing DC works for EI transformers. Cancelling the electrostatic leakage is much more difficult though.

analog_sa said:

And last question. Is it possible to use split bobins (on EI)?


Of course.

analog_sa said:

Would the reduction in capacitance between primary and secondary be beneficial?

Maybe. This will depend upon a few other issues as well, with EI Transformers often the leakage to the core (which has to be then grounded) causes the larger share of the problem.

Sayonara
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.