Attachments
Jac and Jack, thank you for your input and suggestions.
I started listening with the Impasse as stock, tubes are EH for both positions. To my ear playing through the pair of F4s, the sound was gritty almost too grating to listen to for very long. I switched back to my BA-3 pre and much smoother/rounder and a fuller sound.
I have in hand a CBS 6SN7GT along with an Amperex 6DJ8 "Bugle Boy" made in Holland. So those replaced the EH tubes and I took Jack's suggestion and add a 10uf cap across R8 in the power supply.
Sounds better, closer to the BA-3 pre sound. Not gritty or coarse sounding.
Digikey order arrived today with those 1w SMD resistors. So Jac I replaced the R4, R11, R12 resistors with the quad array SMD resistors per your suggestion.
Now the Impasse is really singing through the F4s. Much smoother sound. Listening to "Take Five", really nice, the clarinet is clear and defined. The bass sounds tighter also.
I have no way to make any measurements, just my ears opinion!!
Sy created a great little pre and it plays well with Nelson's F4s.
Tnx
David
I started listening with the Impasse as stock, tubes are EH for both positions. To my ear playing through the pair of F4s, the sound was gritty almost too grating to listen to for very long. I switched back to my BA-3 pre and much smoother/rounder and a fuller sound.
I have in hand a CBS 6SN7GT along with an Amperex 6DJ8 "Bugle Boy" made in Holland. So those replaced the EH tubes and I took Jack's suggestion and add a 10uf cap across R8 in the power supply.
Sounds better, closer to the BA-3 pre sound. Not gritty or coarse sounding.
Digikey order arrived today with those 1w SMD resistors. So Jac I replaced the R4, R11, R12 resistors with the quad array SMD resistors per your suggestion.
Now the Impasse is really singing through the F4s. Much smoother sound. Listening to "Take Five", really nice, the clarinet is clear and defined. The bass sounds tighter also.
I have no way to make any measurements, just my ears opinion!!
Sy created a great little pre and it plays well with Nelson's F4s.
Tnx
David
Last edited:
David,
I don't know for sure. You have gone as far as I have so far. Let's put it this way, quad resistors theoretically have lower distortion and noise than a comparable wattage metal-ox or thick film. Where you have a power resistor that affects the signal, a quad resistor might be worth a try.
Where you already are using a metal film resistor because you don't need high power, my guess is that it would not be worth the trip. But I've never tried it.
Regarding power resistors, I am planning on trying quad resistors in the F4 for the 0.47R resistor locations (R16 - R21). There is a Susumu current sense resistor that looks very interesting for this. It's 1 W, foil resistance element, and available in 0.47R. Mouser part # 754-KRL2012EMR470FT5 , if you are curious.
Jac
Jack,
I think you mean Panasonic ERG power resistors. I'm not one to doubt NP, and maybe his use means that things like current noise are not important. It is true that the power density of current noise is much lower than the signal. That said, the 2009 study of current noise by Frank Seifert shows the ERG to be much noisier than a good metal film resistor. For example, per the Seifert study, at 20 Hz, the current noise density of the ERG is more than 10 times that of a Dale CMF55 and the Dale isn't a particularly quiet resistor. Comparing to something like the Welwyn RC55Y or a Susumu RG, the ERG is more than 100 times more noise.
It may be that this noise isn't significant in many circuits. But the advantage of the quad resistors is that it allows you to use lower power metal film resistors in an array that spreads out the power. It's just a personal opinion, but I feel that the difference is audible and that quad resistors offer a sound quality improvement for very little more money and effort over all of the power resistors I have tried.
Jac
Agreed. By the time I get around to trying, I bet I will have talked myself into have some small boards printed. 24 would be a lot to hand cut.
Jac
Of course, you are correct. Yet over the years we have moved from a noise floor of -60 or -70 dB to -100 or even lower and heard sound quality improvements at the lower levels. Music contains quiet passages much closer to the noise floor, so when an approach like quad resistors comes along, I don't just say that resistor noise is lower than the -90 dB of the tube and forget about it. I am curious to try it and see if it's worth the trip.
In the past, we all used carbon composite resistors for everything. Now we try to use metal film resistors when ever we can because they have better specs and usually sound better. Just think of this is a way to use metal film resistors in a power resistor where it wasn't possible before.
Jac
A 6922 has noise of 80 to 100nV/RtHz @1kHz
The noise of the Maida LM317 regulator is ~160nV/Rt Hz
Shouldn't have said "irrelevent" -- it's one standard deviation.
Last edited:
Interesting. Per the Seifert study, a 100R, 2W Panasonic ERG is about 210 nV/Rt Hz. Of course, that is a 1/F noise, so it decreases with increasing frequency, but it's still double Johnson noise at 3 kHz. In comparison, that Dale CMF55 is about 20 nV/Rt Hz at 20 Hz. We don't have a measurement of the Susumu that David and I used for quad resistors, but the Susumu RG that is measured is among the lowest noise.
Jac
Send me the part number from DK or Mouser etc and I will solder it up and compare. I can isolate out the power supply noise with a high voltage pile and DC filament supply.
Jack,
I wonder it will be interesting to compare one of this per Sy BOM as 1W parts:
CPF115K000FEE14 Vishay / Dale | Mouser
to parallel/series 4 of these as a 4W parts. I wonder a oversized resistor for the same current will make less noise. I am happy to contribute a few bucks to the experiment.
Thanks,
I wonder it will be interesting to compare one of this per Sy BOM as 1W parts:
CPF115K000FEE14 Vishay / Dale | Mouser
to parallel/series 4 of these as a 4W parts. I wonder a oversized resistor for the same current will make less noise. I am happy to contribute a few bucks to the experiment.
Thanks,