Hard work and much listening sessions pays..... Have been trying to get some RelCap 15n with values =< to 15n and it proved not possible.... it seems I used the last from my good supplier, so I decided to use K71-7 styroflexes 1% NOS and have been fighting with some insistent brightness.... I was using 15.03nF // 47pF and fought a lot with tunning caps in the psu, and cart loading ... no go.... so I finally experimented shunting the 15n with some other styros I got.... 15.03 // 47 = bright.... 15.03 // 47 // 47 = dull.... 15.03 // 47 // 22p = almost perfect.... 15.03 // 67pF = as good as can be done with styroflex
This once more proves that work can convey good results once you are really dedicated to it.... and also that 20pF can really make a difference.
Please take this with a little grain of salt as it all depends on your system and also because all reported values are according to my DMM that might not be the last word in precision... that is why I insist that the ultimate tool is your hears (and other listener's ears).
Also it became very clear to me that all reasonably good dielectric caps (even not boutique ones) can be used depending on your values choices..... Next time I will experiment with wima FKP and see how far can I go
Really revealing.
This once more proves that work can convey good results once you are really dedicated to it.... and also that 20pF can really make a difference.
Please take this with a little grain of salt as it all depends on your system and also because all reported values are according to my DMM that might not be the last word in precision... that is why I insist that the ultimate tool is your hears (and other listener's ears).
Also it became very clear to me that all reasonably good dielectric caps (even not boutique ones) can be used depending on your values choices..... Next time I will experiment with wima FKP and see how far can I go
Really revealing.
Last edited:
Not exact LCRs usually show a bit less value than more. By using 4 wire probes 0.25% LCR I got 15n2 on both channels. After measuring with 1% calibrated generator with non expired certificate I found the curve +/- 0.15dB 100Hz-15kHz. 100pF can make a measurable difference of 0.5dB brighter or duller than textbook.
Micas can sound good but lack some mid body, Styros can sound good but care must be taken to avoid a stark edgy sound.... Teflon on alu can be really extended in the highs but really lack body.... styros on tin foil proved wonderfully forgiving, detailed and full body sounding but now very difficult to get..... I mean that the exact ultimate value depends slightly on the values you use and 20p are enough to make a difference.... this is how revealing the folded can be.
Hi Salas, what do you think about BF862 ? Can it be used for input stage ? .... as k369 or k363 are now really difficult to find, maybe I could try these SMD... just need to make a suitable adapter for my p2p builds
I have about 50 of those, they don't come in grades. Most will show over 15mA in not so close samples batching.
Difficulties in this circuit:
-PITA to select IDSS due to SMD. They also need cooling surface when doing so.
-Larger Rs to keep them cool in circuit=extra noise. Need run lower Ibias. Can't afford less Vbias in this FC.
-Needing double the K369s in 862s lot for ~same wide band noise spec.
-Made for radio thus wide BW & easier to oscillate, layout is critical.
Generally:
-Available from main vendors.
-Better than the K170 for noise
-Worse than the K369 especially for Generation-Recombination noise in LF
-Best suited for unselected use in gain fixed headamp with loop NFB.
Hi Salas,
I have one theoretical question about folded and similar phonos assembles parts.
You recommended to use RG174 coax for signal wires. It also probable good for Raw DC in.
Now, do you connect central wire of coax to the signal "+" and braid to signal "-"?
I heard that some people uses only central wire to connect all signal point and braid is connected to ground on one end only to avoid a loop.
Please let me know.
I also have a question about RG179 coax. Is it a good choice for Phono and MC?
I have plenty of it, and I do not have GR174. I thought I might use it instead.
Thank you.
I have one theoretical question about folded and similar phonos assembles parts.
You recommended to use RG174 coax for signal wires. It also probable good for Raw DC in.
Now, do you connect central wire of coax to the signal "+" and braid to signal "-"?
I heard that some people uses only central wire to connect all signal point and braid is connected to ground on one end only to avoid a loop.
Please let me know.
I also have a question about RG179 coax. Is it a good choice for Phono and MC?
I have plenty of it, and I do not have GR174. I thought I might use it instead.
Thank you.
RG174 is not adequate for best carrying DC PSU lines.
Only for signal. RG179 is good for signal too.
For signal use each shield's end should go to respective GND point.
Loops should be avoided in the general grounding, not by compromising the signal's shielding.
For DC you can either use twisted pair wiring of healthy gauge or XLR pro cable. In this case there is a shield, just ground it's one end to PSU 0 or at chassis near the entry point.
The only things I care for about those signal cables regarding the audio band its they are thin, flexible, well shielded and widely available.
Their impedance differences will be interesting only in RF and instrumentation in long lengths. Use either. If you want audio marketed types, keep them short, thin to be flexible, well shielded also. Me I used 174 that I had and I am happy. Now if I tell you its copper clad STEEL core cable you may shrug or look for oxygen free pure copper alternatives.
Their impedance differences will be interesting only in RF and instrumentation in long lengths. Use either. If you want audio marketed types, keep them short, thin to be flexible, well shielded also. Me I used 174 that I had and I am happy. Now if I tell you its copper clad STEEL core cable you may shrug or look for oxygen free pure copper alternatives.
Each input is a circuit. That circuit MUST have a Flow and a Return.
That Flow is commonly referred to as Signal, or Hot.
That Return is commonly referred to as Ground (very ambiguous), or Signal Ground, or Cold (generally only in balanced signalling).
The connection from the cartridge to the RIAA Pre must be a two wire connection to form that circuit. If you use a shield as the Return part of that circuit you MUST connect the shield at the cartridge end and at the RIAA Pre end.
If you use a twisted pair for the connection then again it is a two wire connection. Both the Flow and Return must be connected at both ends to complete the circuit.
If you use a shielded twisted pair, or shielded star quad (microphone cable), again these are part of that two wire connection. The Flow and Return are connected at both ends. The Shield is connected to Chassis at the point where the cable enters the Chassis, NOT at the PCB inside the Chassis. If there is a Chassis around the cartridge, or cartridge wiring, then the shield should be connected to that enclosing metalwork, eg the arm tube and headshell.
Thank you again for such detailed info.
How about RCA connectors returns? Do you connect them together at the MM OUT?
if one connects the returns at the Source and one also has the returns connected somewhere downstream, then one has created a very long loop. That loop is very likely to have problems with picking up interference.
Recently I can across a dual coax with the screens in contact. That would be one way to minimise the loop area of that long loop.
I have never seen any project describe how to avoid this stereo source to stereo receiver loop.
There are a couple of papers that describe methods of reducing the current in these loops.
Recently I can across a dual coax with the screens in contact. That would be one way to minimise the loop area of that long loop.
I have never seen any project describe how to avoid this stereo source to stereo receiver loop.
There are a couple of papers that describe methods of reducing the current in these loops.
Last edited: