Really strange... I can't figure out how the amp could be damaged in this way...
You don't have any socket or similar, right?
Double check cabling and his polarity, if you reverse input hot and ground you can have a large amount of hum (but in this case DC offset is normal).
Your layout is still this one?
The boards were fixed to the bottom plate by screws, right?
Last edited:
Hi Dario.
Layout is the same. Only twisted wires. Hum, as I understood is ok with disconnected amp. The other thing is b1 which was not really housed and Has obbligatos loosely attached(put on the own wires). I. Guess it Could cause the noise. I did Not Hear noise with No pre attached, only hum.
still values of dc offset differ with inputs shorted and right channel is swinging up down to Negative Badly...
Layout is the same. Only twisted wires. Hum, as I understood is ok with disconnected amp. The other thing is b1 which was not really housed and Has obbligatos loosely attached(put on the own wires). I. Guess it Could cause the noise. I did Not Hear noise with No pre attached, only hum.
still values of dc offset differ with inputs shorted and right channel is swinging up down to Negative Badly...
Will, Didn't you have hum problems with the metal shell of the Audyns being too close to the chassis ground? Ivan has just some thin double sided tape. I also see the cap lead being right next to some other components. Don't know if that routing (antenna effect of cap lead loop) is causing a problem.
Attachments
Yes, I had some problems with the metal body of the k71-4 interacting with the mounts that I made for them. Some of the other caps that I used also had similar effects. In my situation though I was only able to eliminate the hum by using the pgnd connection.
On a side note I have been auditioning the true coppers and have really been enjoying them. Following the advice of madisonears I have not been switching caps as often as I was, so the TCs have gotten an extended listen through a variety of genres. Just this morning my five year old and I were enjoying some essential Dave Brubeck. The next cap I will be auditioning will be the jantzen silver z, if I ever pull the TCs out of the system
We need to figure out a way to make the TCs easier for north american builders to squire.
On a side note I have been auditioning the true coppers and have really been enjoying them. Following the advice of madisonears I have not been switching caps as often as I was, so the TCs have gotten an extended listen through a variety of genres. Just this morning my five year old and I were enjoying some essential Dave Brubeck. The next cap I will be auditioning will be the jantzen silver z, if I ever pull the TCs out of the system
We need to figure out a way to make the TCs easier for north american builders to squire.
I have disconnected left board to check soldering- all fine. Put it back and my (not autoranging) mm on 200mv scale showed 0.44 or similar.
So left is fine- I have had something connected wrong ( !) , and btw I have opened up a bit to be able to put it easier (for now). It could be the case also
But left channel, as it was before shows anything between 0.4 and 0.7 and never stabilizes..
Inputs were shorted .
So left is fine- I have had something connected wrong ( !) , and btw I have opened up a bit to be able to put it easier (for now). It could be the case also
But left channel, as it was before shows anything between 0.4 and 0.7 and never stabilizes..
Inputs were shorted .
Attachments
So DC offset is 0.44 mV?
Incredibly low... if so you can try connect test speakers and tell us if you hear music... I bet it's not the case....
The TCs lead extension was insulated when the incident happened?
If not it's possible that you had high DC on LM318s and fried them.
It would be nice if you measure DC voltages on regulator.
Suggestions
Hi.
Now I am starting to upgrade from (almost) standard Mouser BOM to better components.
I have put sockets in the PCB so I can change the capacitors/resistors the Dario way - good idea!
I am still waiting for Black Gates as C9 and I have not yet experimented with the small capacitors (C7, C12, C30, C32) but I have found some other things to do , witch enhances the performance, IMHO.
First I placed two 10uF/25V Solid Alu Tantal as C101/201 to hear what they could do . They are old favorite of mine for decoupling. And - yes - better focus and better instrument separation. Not a lot though.
Next I replaced the Caddok 50 Ohm R101/102 with another favorite of mine, Rhopoint 0.1% wire wound. This is also a favorite of Tosten Løsch and John Brown (ec-design) .
This is highly recomented. It is the same areas the improvements are as the capacitors, just more so and SSS sounds on voices sounds more natural and relaxed. Easier to hear the size of the room and separate the direct sound from the instruments and the reflexions from the room. I tried to remove the capacitors after the resistors were changed, but they went in again as the improvement were much easier to hear after the Rhopoint was in the circuit.
If anyone tries these suggestions, please report..
Hi.
Now I am starting to upgrade from (almost) standard Mouser BOM to better components.
I have put sockets in the PCB so I can change the capacitors/resistors the Dario way - good idea!
I am still waiting for Black Gates as C9 and I have not yet experimented with the small capacitors (C7, C12, C30, C32) but I have found some other things to do , witch enhances the performance, IMHO.
First I placed two 10uF/25V Solid Alu Tantal as C101/201 to hear what they could do . They are old favorite of mine for decoupling. And - yes - better focus and better instrument separation. Not a lot though.
Next I replaced the Caddok 50 Ohm R101/102 with another favorite of mine, Rhopoint 0.1% wire wound. This is also a favorite of Tosten Løsch and John Brown (ec-design) .
This is highly recomented. It is the same areas the improvements are as the capacitors, just more so and SSS sounds on voices sounds more natural and relaxed. Easier to hear the size of the room and separate the direct sound from the instruments and the reflexions from the room. I tried to remove the capacitors after the resistors were changed, but they went in again as the improvement were much easier to hear after the Rhopoint was in the circuit.
If anyone tries these suggestions, please report..
The solid alu capacitors are Phillips SAL.
Farnell has Rhopoint:
8G16D 50R 0.1 - RHOPOINT - RESISTOR, 0.33W 0.1% 50R | Farnell United Kingdom
Farnell has Rhopoint:
8G16D 50R 0.1 - RHOPOINT - RESISTOR, 0.33W 0.1% 50R | Farnell United Kingdom
On a Rev C, I use styroflex on all of the above, with small gains over FKP2s overall. It does become more revealing and analytical, so it may not please everybody.
C7 - ERO KP1834/35
C12 - Siemens/generic 220pF axial
C30 - K71-7 1000pF (a bit large, but fits in V1.2..V1.4 board)
C32 - ERO KP1834/35
Anything that makes voiced sibilance more natural is welcome. I'll probably try Shallcross RWR80 or similar wirewounds when I try the LM318 CCS mods to the Rev C.
I've already posted about FKP3 (instead of MKP10) for C4 on another thread - it's the best I've heard at that location, but again more analytical and revealing.
Awaiting Nichicon KZ 220uF/100V for C1/C2 instead of the 50V caps I've been using so far - the 100V caps have higher Ir, lower ESR and lower tan-delta than the 50V caps, and they're the same 16mm diameter. I've tried Elna RE3 220uF/100V at C1/C2, and they're very good, but with magnetic leads. Also tried Black Gate F for C1/C2, but the improvement over KZ is not that great to justify the cost - but maybe I'll change my mind after it runs in a few 100 hours.
This may indicate that decoupling is not yet optimised.
That would be decoupling of both active sections, requiring at least 8 decoupling capacitors.
I have some ERO KP1838, wich I think is the same as KP1834 but axial instead of radial.
They are 100pF and 4,7 nf.
Will the 100pf be ok as C32 or will there be stability issues.
The 4,7 can be used alone or as two for C7.
I also have the same values in Suflex, also an old favorite of mine. I used these in tube riaas I manufactored in mid ´80 before I had to change to ERO KP, as Suflex became very hard to get.
I will try to get som FKP3 for C4. I have always found MKP10 a little soft and muddy (better than MKP4 though) I always preffered RIFA PHE 431. They are also more analytical and revealing and , as I remember , could be a little too much in the top spectra.
Is FKP2 as good? Cannot get FKP3 and it looks like FKP2 is the same , but in 5mm raster instead of 7,5 or 15?
Please report back on the Ellyts for C1/C2.
They are 100pF and 4,7 nf.
Will the 100pf be ok as C32 or will there be stability issues.
The 4,7 can be used alone or as two for C7.
I also have the same values in Suflex, also an old favorite of mine. I used these in tube riaas I manufactored in mid ´80 before I had to change to ERO KP, as Suflex became very hard to get.
I will try to get som FKP3 for C4. I have always found MKP10 a little soft and muddy (better than MKP4 though) I always preffered RIFA PHE 431. They are also more analytical and revealing and , as I remember , could be a little too much in the top spectra.
Is FKP2 as good? Cannot get FKP3 and it looks like FKP2 is the same , but in 5mm raster instead of 7,5 or 15?
Please report back on the Ellyts for C1/C2.
Last edited:
Thanks, I'm happy that you're experimenting
Mmh... Even compared to Black Gate PKs (the best cap I've heard and tried in C101/C201) I've alway preferred to leave void those positions.
Yes, with caps it seem to sound more focused but not as natural sounding as without, IMHO.
Those Rhopoint must be really good resistors to beat Caddocks...
BTW are you sure that the improved detail you're earing is not due to a different timbre (towards HF)?
The bass response changed?
Hi Andrew,
we already discussed in the past about it... maybe the LM3886 decoupling is not 'optimized' but is the one designed by Penasa...and it sound best to my ears Vs the classic 4 caps decoupling to ground.
It could be different for the LM318 regulator, though, since I've chosen to not decouple low freq. at all, HF decoupling should be addressed, as for the LM3886, by the rail to rail cap.
Having the right instrumentation it would be nice to investigate the latter, sadly I don't have it and I have to rely (mostly) on my ears only.
No, 100pF is not enough, you may have stability issues.
Also, 4.7nF is probaby not enough for C7 and I would avoid paralleling them, it would probaby lead to longer leads and added inductance.
if changing the capacitor changes the sound, then at least one of the implementations is not amplifying correctly.
That conclusion cannot be changed just because it has been discussed previously.
Whether Penasa did or did not optimise his implementation does not have significant bearing on whether the current implementations are optimised for decoupling.
In my view, removing some of the decoupling cannot be better than optimising the decoupling for BOTH active stages.
That conclusion cannot be changed just because it has been discussed previously.
Whether Penasa did or did not optimise his implementation does not have significant bearing on whether the current implementations are optimised for decoupling.
In my view, removing some of the decoupling cannot be better than optimising the decoupling for BOTH active stages.
The difference could be that these are relatively small capacitors and the Phillips SLA have very good HF response. I do not think it changes the sound per se.
They ARE very good. I have tried a lot of resistors (including homemade honeycomb wired) as I/V resistor in various non oversampling DACs and I prefer these to anything.
It has nothing to do with rising hf response, if anything the highs are less prominent (the SSS reduction). This is the only thing I hear that could be negative, maybe there is a little loss of air, but the room is more evident, and the details are still there.
I see I make some controversy here, and I do not want to upset anybody. Please try it to hear if you hear the same things or correct me if I am wrong. For now I will leave it here and not press on with my findings.
At least 150pF is needed for stability at C32. 330pF is the max practical value, and requires other changes to retain stability, and also greatly changes the sonics.
Yup, exactly right - FKP3 is cleaner, more defined, more airy than MKP10. I have tried Rifa PHE426, which IIRC was a bit harsh (but it wasn't an exact A/B replacement test, so other things may have caused the harshness). Also tried MKP2, which sounded about the same as MKP10 (maybe slightly less detail). Rifa PHE840 was also comparable to MKP10.
FKP2 is limited to 33nF max, IIRC. If there are larger values, go ahead and try it out.
I'm not sure about this, but I think FKP2 is stacked film/foil, while FKP3 is wound film/foil. Again, not sure which construction would be better in this application. FKP3 tends to be larger, which might help the pulsed current rating.
I don't agree but, obviously, this is not the point.
Most readers didn't followed all the My_Ref threads so I've simply stated that we already discussed about it, so that is clear that it's not a new 'problem' but simply a design choice we disagree about.
Sure that those capacitors changes sound... try another pair from a different manifacturer, I bet it will sound different.
Just to be clear...
Rhopoint resistors you're talking about are these?
An externally hosted image should be here but it was not working when we last tested it.
If so the one you linked seem a very different beasts.
Fine, but bass resposnse has changed?
No controversy at all, absolutely don't give up and keep us informed... all experiment reports are welcome