X-Can v3 - modded vs non-modded - blind tests

[Deleted]

ah... 'drie-en-dertig piepschuim condensatortjes met een zeer korte brandpunt afstand'

[szoze]

Apr 24, 2012 10:20:24 GMT @solderdude said:

Coin flippin'....

I flipped some coins to get a better statistical view on these test where the 'boundaries' lie of what can be attributed by chance.
If the differences are obvious/night and day one would expect 11 or 12 out of 12 tries, perhaps when less audible 10 out of 12 would be correct.

I completely agree with you Frans. The difference between the amps was far away from night and day.

[szoze]

Apr 24, 2012 13:33:57 GMT @solderdude said:

ah... 'drie-en-dertig piepschuim condensatortjes meteen zeer korte brandpunt afstand'

I am also a little surprised that Mike uses such a small capacitor. Personally I would guess something in nF range, maybe up to 22 nF

[Deleted]

Ah he doesn't.

As far as can be seen in his 'work in progress' thread he only bypassed the diodes once using 10nF ceramic multilayer caps.
If he pulls your leg any further it might even come off. ;D
???I wonder who can come up with the closest correct translation of my (witty) reply above...

1nF seems a sensible value that can suppress the very small HF artifacts in the secondary transformer coil.

For getting rid of differential HF noise that is not filtered out by the reservoir cap (because the freq is too high, i.e. above 10MHz) a 10nF to 100nF cap directly over the PS cap should do the trick.

[szoze]

Apr 24, 2012 13:47:36 GMT @solderdude said:

Ah he doesn't.

As far as can be seen in his 'work in progress' thread he only bypassed the diodes once using 10nF ceramic multilayer caps.
If he pulls your leg any further it might even come off. ;D
???I wonder who can come up with the closest correct translation of my (witty) reply above...

1nF seems a sensible value that can suppress the very small HF artifacts in the secondary transformer coil.

For getting rid of differential HF noise that is not filtered out by the reservoir cap (because the freq is too high, i.e. above 10MHz) a 10nF to 100nF cap directly over the PS cap should do the trick.

Thanks Frans. It will be very hard to take Mike seriously after this.

I am not sure I can translate your message word by word but I can certainly understand it's core

[szoze]

Apr 24, 2012 13:47:36 GMT @solderdude said:

1nF seems a sensible value that can suppress the very small HF artifacts in the secondary transformer coil.

But, how is the larger forward voltage drop on the UF4007 diode going to influence the circuit?

[Deleted]

In this particular application (unregulated power supply) the power supply voltage will be 0.2V (1%) lower than when a UF4001, UF4002 or UF4003 were used and about 0.1V lower compared to 1N4001-4007 (0.5%)

effectively this means with UF4007 you will have 0.08dB less dynamic range compared to the original diode.
With UF4001-4003 you will gain 0.08dB headroom.

listening tests have shown even in a direct comparison 0.1 dB cannot be detected so in maximum output LEVEL the differences of 0.08dB wouldn't be noticeable to the majority of the human population.

When you replace them in a circuit where the reservoir cap is followed by a regulator or SMPS there will be no difference in power supply voltage or noise levels.

In switching applications (used as a snubber over a switched inductor/trafo) it would mean there would be higher RF emission compared to UF4001-3

There are also differences in on-off behaviour between 1N4001 and 1N4007 by the way due to their physical construction.
It's not only the reverse voltage that differs (same for UF4001 and UF4007)

The sonic implications are not my department and also have not seen reports about differences between 1N4001 and 1N4007 although on-off speed differs considerable between the 2 types.

Experiment... if it brings SQ improvements to your ears I feel it is worth doing (doesn't cost much either), if it doesn't you will at least know you have brand new diodes in there for peace of mind about aging.

P.S. the translation of my question above (or closest to it) would be:
33 squeeky-foam capacitors with a very short focal distance.

[FauDrei]

Thanks for detailed "cap over rectifier diode" explanation Frans.

That, if I understood you correctly, means my headamp diode swap for SIC Schottkies was, more or less, to appease my mind... Oh, well...

A question: I have 10nF film snubbers over them (they were there even before diode swap). What would from technincal perspective change if I remove those snubber caps?

[Deleted]

The difference between the 2 diodes might become evident when used in the intended applications.
This means Switch Mode Power Supplies, power inverters (= SMPS) and high power motor controls where high reverse current peaks need limiting to protect the switching devices from a premature death due to extremely high reverse voltages generated by the inductors/motors/transformers when the current is switched off (10,000 to 100,000 times per second).
That is where these special diodes are designed for, it cannot be done with 'normal diodes'.
Interchanging these diodes may reveal slight differences in RF emission from this kind of equipment or have influence of the efficiency (thus heat emission).
You wouldn't be able to use them in those applications anymore when you would fit even 100pF across them though.

For rectifying mains they are both kind of overkill and add nothing, as explained earlier the difference in nanoseconds is negliable small compared to the actual time in ns the diodes are conducting in 50/60Hz rectifiers.
The MUR420 has the edge in forward voltage drop (around 0.15V lower at around 0.5A) over the C3D06065A
The C3D06065A can handle slightly more current though.

The snubbers across the diodes serve 2 purposes both of which are HF (emission) related.
For the DC voltage on the reservoir caps they have no influence.

Effectively what the capacitors do is 'short' any spurious HF entering FROM the trafo as well as 'snubber' which is kind of 'smoothing the transition point' the diodes.
Relatively spoken the voltage change (at the switching point) is so enormously small (in the mentioned ns) that the diodes (even the slow ones) have no trouble keeping up with it.

HF garbage from the mains is not coming through a trafo by inductance but comes from capacitive coupling of the primary and secondary windings.
Toroids are worst in that aspect followed by E-core and R core has the least HF coupling (in both directions).
Also there are VERY small HF peaks on the secondary side of the trafo (couldn't see them on the scope but sensitive RFI receivers may pick up very low level RF) at the moment the diodes start and stop conducting.
They conduct only a fraction of the time (when the sinewave of the trafo reaches the voltage that is in the reservoir caps) and the rest of the time they are in 'reverse' and do not conduct.

anyway... the capacitors effectively form a 'short' for HF signals (above 1MHz) between the 2 secondary windings as well as between the + and - of the reservoir cap.
effectively almost the same mounting a 5nF cap across the secondary windings and one 5nF over the reservoir cap.
There is a difference though as it also prevents the diodes from emitting 'switching' artifacts, which are extremely low amplitude and high in frequency in this case anyway (when not snubbered).

So if your reservoir cap already has a 10nF (or even 100nF) decoupling cap across it already you can omit the 10nF but for these low frequencies (100Hz sinewave and perhaps) removing those caps will go unnoticed.
No reason to take the 10nF caps out though, but mounting a very fast diode and making it extremely sluggish again defeats any profit that can be had 'speed' wise from the diodes.
The differences in forward voltage also aren't very shocking.
You must realise the diodes only 'connect' the mains (via the trafo) around 1ms every 20ms the rest of the time it is 'off' and mains is litterally disconnected by the diodes.
So the general idea in hifi land that audio currents come directly from mains at all time is hereby declared nonsense as well as most of the time (90 to 95%) the mains aren't even supplying the power at all... the reservoir caps do and get charged every 10ms (or 20ms) for a short period of time.
Only for HF garbage there is a continuos connecting with mains when snubber diodes are fitted, but at the same time is 'shorted'.

So technically the change of diodes (certainly with snubber caps) made no technical difference in THIS application except for the slight loss of voltage on the reservoir caps.
When regulators are used (in most cases) this would go unnoticed.
???How removing the caps and changing the diodes alters the PERCEPTION of the music is an entirely different matter and is not technically explainable by the speed or other technical factors (that are known to the technical people).

PS I know I am completely ignoring the thousands of reported improvements that appear to prove me wrong.
I don't mind I am free to speak my mind (I hope) and if people clearly hear differences than at least to them it is a very worthwhile upgrade.
I am completely O.K. with that b.t.w.

[FauDrei]

OK, all mentioned caps I have... so all I did was a nice (de)soldering lesson. ;D Nevermind.

Thanks for further explanation. Can you also explain why those TO220 SIC Schottkies get hotter than original "normal shaped" diodes even though their current specs are higher than originals? Is this because, as you explained, majority of time they are just blocking?

[Deleted]

A question: I have 10nF film snubbers over them (they were there even before diode swap). What would from technincal perspective change if I remove those snubber caps?

Valter
To hell with "technical" explanations in a highly subjective area like this. ;D
Try them without, and if they sound the same or better to your ears, leave them off. (KISS principle.)
If in the unlikely event it sounds worse, put them back on!
Kind Regards
Alex

[Deleted]

Can you also explain why those TO220 SIC Schottkies get hotter than original "normal shaped" diodes even though their current specs are higher than originals? Is this because, as you explained, majority of time they are just blocking?

When diodes 'block' they do not conduct so dissipate no heat.
Regard it more like a switch that only conducts in 1 direction but in that direction forms a small (not linear) resistor.

The reason for the difference in temperature is the higher forward voltage and physical construction.
The forward voltage is the voltage that falls across the part in conductive state and depending on construction/type/current can be anything between 0.3V and almost 2V.

The TO220 diode has a higher voltage across it when conducting and with the same current this means a higher wattage and thus more heat.
For that reason they are TO220 so they can (and for higher currents must) be bolted onto heatsinks in which case they can handle more current then the normal counterpart that doesn't have the heatsink possibility and relies on conduction of heat through the wires and convection (forced or natural).

During the short time a diode conducts it needs to recharge all the 'power' (charge in Coulomb) drawn from the cap during the relatively large period the diodes are not conducting.
So, if for instance the circuit draws 1A continuously, during the time the diodes are 'recharging' have to handle a current that is needed to recharge the cap again in a MUCH shorter time period.
This means that eventhough the 'average' current drawn from the secondary is 1A (simplified, the actual effective current differs) and that's what an amp meter will show (the average current, not the momentarly peaks) in reality the current in the diodes is many times higher.
The heat in the juction itself rises high in that short period and cools off in the 'unused' time.
Because this happens fast (50 times per second) the part itself 'smears' the heat over time.

In the case of the TO220 diodes the curve is less 'steep' compared to the other diode.
This means the voltage drop is even higher as when used as a diode in DC paths.
Therefore the power dissipated is higher as well and thus get's hotter than the other one.

The higher the value of the reservoir cap the higher this current as also the conducting time is shortened as well.
These high current peaks are often the reason why some diodes suddenly fail after a long period of usage in (bridge) rectifiers in higher power applications where big reservoir caps are used.

[szoze]

OK, we have now compared two x-cans, one modded and one non-modded. They are identical in every detail except for the rectifier diodes. One of them has the stock diodes and the other one has UF4007 with 1 nF snubbers. My two friends and I did a sighted AB test.

We listened and listened for hours directly comparing these two amps and there is simply no difference to be "heard". The amps sound, in our ears exactly the same. We are ready to give 1000 Euro award to anybody who can objectively demonstarate (proove) that there is any difference in sound quality between these two amps.

[gommer]

When diodes 'block' they do not conduct so dissipate no heat.

Hehe, try putting 400V reverse voltage across a suitable diode and see what happens. Heat it and you'll have thermal runaway, only with static reverse voltage.

Agreed, not applicable in audio land, just warning in case anyone wants to try

By teh way, i would translate 'drie-en-dertig piepschuim condensatortjes met een zeer korte brandpunt afstand'
as
33 tiny EPS caps with a very short focal distance.

[PinkFloyd]

Apr 27, 2012 17:53:11 GMT szoze said:

OK, we have now compared two x-cans, one modded and one non-modded. They are identical in every detail except for the rectifier diodes. One of them has the stock diodes and the other one has UF4007 with 1 nF snubbers. My two friends and I did a sighted AB test.

We listened and listened for hours directly comparing these two amps and there is simply no difference to be "heard". The amps sound, in our ears exactly the same. We are ready to give 1000 Euro award to anybody who can objectively demonstarate (proove) that there is any difference in sound quality between these two amps.

Let me understand..... you have TWO stock X-CAN V2 amps and in one you have fitted 6 x UF4007 with 1nF snubbers? ie: BOTH amps totally unmodded apart from the six diodes in one of them?

If that's the case I'd tend to agree with you, you will hear little, if "any", difference doing it that way round... the diodes were the "last" thing I experimented with when going through all the mods... firstly you have to get the clarity of the amp up to it's max potential and "then" you may be able to detect the subtle improvements the faster diodes bring about.

Remember, every little tweak will be subtle (not major)... if it's a major difference you've connected something up the wrong way round

The idea is that 20 "subtle" improvements will equate to ONE major improvement, this is why it takes time going through the entire amp noting every improvement / no change / worse.....

Also remember that if you are listening to two "stock" X-CAN V2's you are listening to two amps that probably have dried out caps / out of spec valves.... I listened to one the other day (a customer's amp in stock form) and it was absolutely "dull"... replacing all the caps improved it LEAPS AND BOUNDS, in every area.

Now, maybe my "mods" aren't so much mods but more of a "renewal" of a tired amp to it's former glory... whatever the case may be a tired old V2 sounds lifeless compared to one which has been treated to a new set of caps / valves.

It's a sweatshop in that can and the 85C general purpose caps will NOT live forever.... sure, the amp may sound ok but it's like driving with a noisy exhaust, you only really become aware of it when it drops off.

If you WANT to do these tests PROPERLY then get hold of some brand new Jamicon caps, recap both amps (at the same time) allow both to burn in for 100 hours and THEN start tweaking one of them and comparing it to the other... that's the way I did and the results (from what the customers have said) have been well worth the effort.

Mike.

[XTRProf]

Apr 27, 2012 20:01:09 GMT PinkFloyd said:

Now, maybe my "mods" aren't so much mods but more of a "renewal" of a tired amp to it's former glory... whatever the case may be a tired old V2 sounds lifeless compared to one which has been treated to a new set of caps / valves.

I agreed. Just changing the valves for the new Rusky at least got rid of the microphony very obviously in my case. At the same time, it improved clarity especially and all others across the board as well.

Now more of ...........





Then what you think the same play to death video?