Yes, your impression on the C12 mod is what I would expect on normal systems. I think I have already explained why. I agree that change of different caps will provide different sounds, having done much measurements and study on caps, although I have a handle on what to expect just by looking at their measurement data, it is very hard to elaborate, and lots of things do not appear in the general textbooks. I can now understand how Einstein felt when he formed his relativity theory, too bad I am not that good at mathematics.
Have you tried adding the cap parallel to C4, or removing C1 & C2? I realize it's not so convenient to get parts as around here. What is interesting is that the local electronics parts stored here are getting more and more customers.
I wonder how many people feel that the proper polarity when playing back music is significantly important in their systems? Especially with classical, Jazz, and any live performance.
I've started to listen to orchestral music at lower volume. Higher volume still seems a bit messed up. Since I can't really be sure whether it's the room or not, I've located a source for 63V 1000uF electrolitics, seems to just satisfy my +/- 27 V rail. Will report more when I get them. Right now, listening at lower levels with polarity inverted, lots of orchestral recordings sound really good.
Has anyone actually pushed the MyRef to clipping levels?
Has anyone actually pushed the MyRef to clipping levels?
Last edited:
I have said similar many times. Chipamps run out of current ability if the volume is turned up.
This is true with many amplifiers. The question is where is the true limit, what is the cause of the limit. In general, I think the limit would show up like driver thermal compression. Up to now, from an aural point of view, I seem to have pushed the limit up with the mods.
Last edited:
There are some occasions where I start to get noise from the amp. Up to now, it has only occurred when I have the amp upside down, and only occurs in the channel with the original LM318 compensation. I tend to assume it's some form of instability?
Seems like a signal wire had a broken insulation that was touching the chassis occasionally when upside down.
C7 and C4 are better with X2 caps, currently I am still playing around with the capacitance value to see exactly where the right balance is. I can only go as high as the hole will allow. Probably there will be a way to drill bigger holes.
The original caps are AVX BQs (yellow boxes)?
If so a lot of caps are better... 😉
While bigger caps could seem to sound better you can be shure that their effectiveness as decoupling caps is much lower...
When, finally, you'll receive the FE kit with 10nF Wima FKS2s for C7 you'll agree. 😉
I've had even better results with NOS polycarbonate 10nF Wima FKC2.
I think current duty cycle capability is an issue. Larger value caps are bad when they charge and discharge through the ground. I find large caps decouple just fine up to the MHz range, why would you say they are not as effective in the C4/C7 usage?
I have tried Wima 10nF FKS (green) in C7, they provide a very unnatural high frequency and also compresses image depth making the sound louder as well. They are great for cap bypass though. I use them a lot. My current cap for C7 is 0.15uF X2 cap. C4 is 0.22uF X2 cap.
I have tried Wima 10nF FKS (green) in C7, they provide a very unnatural high frequency and also compresses image depth making the sound louder as well. They are great for cap bypass though. I use them a lot. My current cap for C7 is 0.15uF X2 cap. C4 is 0.22uF X2 cap.
Last edited:
Because larger caps means longer leads and traces, so bigger inductance and bigger risk of resonances, isn't it?
Quite in every AppNote on decoupling they stress on it, short leads, as near as possible to power pins...
Well, X and Y caps are designed for very high frequencies, mainly used for mains power decoupling. So I cannot see it being a problem. X caps go as large as 2.2uF. A variety of cap structure and material have different effects and are better for different purposes, one could write a book on it, and I have read a report that was quite thick. App Notes are fine for initial reference, but a lot additional considerations need to be taken into account when it comes to actual interface with the human.
BTY, snubbers are supposed to suppress resonances.
BTY, snubbers are supposed to suppress resonances.
Last edited:
If you are going to order caps, best go for 2200uF 100V (or at least 80 depending on your DC rail voltage). But I was also wondering, the current large power caps on linear power supplied might also perform the same function because they are grounded at the star ground. I have not studied the linear power supply section of this circuit. Maybe, after removal of C1/C2 just put the cap across that location.
Last edited:
When I've read your post and the high value you were suggesting I started thinking exactly the same thing.
Maybe also the switching power supply takes advantage of big smoothing caps.
It would be also useful if, when you try a mod, you specify the exact alternate component used (make/model).
It would be better, when the replacement component have a different value, if you use the same make/model.
If I learned a thing from the My_Ref it's the fact that it let you hear differences beween alternate components of the same value... 😉
I basically use whatever is in the store. For the electrolitics 2200uF 63V I tried was "forever". I don't think this will be of good information because there are just a variety of brands that may or may not be making for other brands. I think if people find the concept satisfactory, then they can determine which brands they like better. Basically 2200uF was the most I could find available around here that would fit into the chassis that I have, 1000uF to 2200uF was not so big a change although it was clearly audible to me, so I am thinking probably I am getting close to the maximum.
In linear power supplies, the big caps also might tend to act like a rail to rail cap of half the size, this is what I remember from some EMI testing that was conducted here. This could explain why what I hear is different from others here. For example, the TP kit comes with two 10000uF caps, so the effective rail to rail capacitance would be something like 5000uF. So it might mean you are starting out with better sound quality than I.
In linear power supplies, the big caps also might tend to act like a rail to rail cap of half the size, this is what I remember from some EMI testing that was conducted here. This could explain why what I hear is different from others here. For example, the TP kit comes with two 10000uF caps, so the effective rail to rail capacitance would be something like 5000uF. So it might mean you are starting out with better sound quality than I.
Last edited:
When capped to ground, the power will not start because the surge current triggers the protection circuitry. I still think it's keeping the ground as constant as possible by balancing current flow though it. I would rethink the layout after all of this. I know there is lots going on at the ground point because that is what's been causing me problems with EMC tests. Disconnect the earth ground and all is okay. So ground is not really ground as we know it.
When I've read your post and the high value you were suggesting I started thinking exactly the same thing.
Maybe also the switching power supply takes advantage of big smoothing caps.
...
That's very high frequency from a 50 or 60 Hz power line point of view. A power amp or op-amp that's trying to oscillate may have a different opinion on what high frequency is.