Do you have schematics for Millenium... By any chances...
For the CRDs, do you know which ones goes in each Borbelly Super Shunt and how many? I have two Super Shunt regs so it would be for those two.
Thanks
Do
I suspect jfet or CRD is lower noise than Lm 334 . But you can trim LM 334 for precise current value ( even add 1n4148 and second resistor for very low temp drift). I prefer to use selected jfet or CRD.
Pinnochio - if you used 2SK170 for your current sources i don't think changing to CRD will get you any improvement as CRD are just jfet current source. You can trim 2SK 170 better than variance of CRD.
Pinnochio - if you used 2SK170 for your current sources i don't think changing to CRD will get you any improvement as CRD are just jfet current source. You can trim 2SK 170 better than variance of CRD.
Uh, ok, but I did a search and couldn't find a reason to replace a CRD (CCD or CLD) with the LM334 + resistor. It would seem to me to be cumbersome to do that on the Jim's Audio boards... sometimes the actual value of constant current is critical, and sometimes it's not. I don't see where it's all that critical in the Kubota regulator circuit.
Waiting to post BOM
tickhpop got part of it. CRDs are fine, the part is not critical as Stanton pointed out.
I will create a BOM for 'everyman's JC-80' in which the emphasis will be lowest cost for onesy twosy diy. CRDs offer little advantage but cost more so they will be listed as the alternate. Some builders will choose the kit of the Kubota but some of the parts will not be used so low cost favors buying just a pair of pcbs (and getting reduced shipping on second board).
I am putting off posting until I have completed one power chain and checked the noise levels to test the claim of "Noise : <0.1mV" under pure resistive load. There are likely better regulators. There are definitely more expensive regulators. It may be that the proposed power train does not resond well to dynamic loads presented by JC-80 but that will only be known by comparing noise levels during operation with that under pure resistive load.
I hope this helps folks understand that I am not interested in building the best solution but rather the best diy compromise. Blowtorch is, was intended to be, and will remain the 'not for everyman' alternative. The Borbely regulator is in use and should offer lower noise at a higher cost. Ditto for Pardo, et. al. and I would love others to measure the noise levels after the cap multipliers but doubt there will be many existing systems torn apart to do this.
tickhpop got part of it. CRDs are fine, the part is not critical as Stanton pointed out.
I will create a BOM for 'everyman's JC-80' in which the emphasis will be lowest cost for onesy twosy diy. CRDs offer little advantage but cost more so they will be listed as the alternate. Some builders will choose the kit of the Kubota but some of the parts will not be used so low cost favors buying just a pair of pcbs (and getting reduced shipping on second board).
I am putting off posting until I have completed one power chain and checked the noise levels to test the claim of "Noise : <0.1mV" under pure resistive load. There are likely better regulators. There are definitely more expensive regulators. It may be that the proposed power train does not resond well to dynamic loads presented by JC-80 but that will only be known by comparing noise levels during operation with that under pure resistive load.
I hope this helps folks understand that I am not interested in building the best solution but rather the best diy compromise. Blowtorch is, was intended to be, and will remain the 'not for everyman' alternative. The Borbely regulator is in use and should offer lower noise at a higher cost. Ditto for Pardo, et. al. and I would love others to measure the noise levels after the cap multipliers but doubt there will be many existing systems torn apart to do this.
Last edited:
Wolfsin - don't know what you think is " dynamic " about JC 80 as a load. It is biased heavily (for a preamp) class A load and therefore fairly constant as a load.
I am counting on John to avoid the 'bug' trap ;-)
@tickhpop -- purely hypothetical. Question then becomes whether 100nv izGoodEnuff?
@tickhpop -- purely hypothetical. Question then becomes whether 100nv izGoodEnuff?
Perhaps what ticknpop meant was that measuring noise with no signal would be as good as under resistive load. Please, if that is the case then let me know so I can skip ahead, build the whole thing and measure later.
JC-80 gain adjustments
Here's some info on the gain of JC-80 board
"The default gain of JC-80 is around 14. The gain can be adjusted by altering the feedback resistors R17 to R20 simultaneously as JC-80 has a symmetry structure. Basically it is a ratio of R18/R30 but there is constraint since when R30 is changed the operating current will be changed too."
This can surely be useful for all of us using this board
Here's some info on the gain of JC-80 board
"The default gain of JC-80 is around 14. The gain can be adjusted by altering the feedback resistors R17 to R20 simultaneously as JC-80 has a symmetry structure. Basically it is a ratio of R18/R30 but there is constraint since when R30 is changed the operating current will be changed too."
This can surely be useful for all of us using this board
I would expect the noise output of a power amplifier to be
The noise from a source component should be at least 12dB and preferably more than 24dB quieter than that, over the same bandwidth, i.e. <25uVac and preferably <6uVac.
i.e. <100uVac, over the 20Hz to 20kHz bandwidth.
The noise from a source component should be at least 12dB and preferably more than 24dB quieter than that, over the same bandwidth, i.e. <25uVac and preferably <6uVac.
Is Kubota quiet enough?
Thank you, Andrew. Now we may be getting somewhere. The Chorbely is advertised as having "Output residue noise <50uV(20Hz-20KHz)" compared to "Noise : <0.1mV" for the Kubota, both falling considerably short of the standard you suggest. And it may be that neither is up to what is needed.
Fortunately the choice of regulator is highly modular in the power chain and thus particularly amenable to testing but if the Kubota can be rejected now, great. We will soon have participants with regulators ranging from genuine Borbely through Chorbely to Kubota. The SuperTeddyReg has repeatedly been suggested but I have no comparative noise measurement for it. Perhaps we should consider a pair of beefy 24vDC batteries as well. Anybody know the noise levels for 1)genuine Borbely, 2)SuperTeddyReg, and 3)truck or solar batteries?
Thank you, Andrew. Now we may be getting somewhere. The Chorbely is advertised as having "Output residue noise <50uV(20Hz-20KHz)" compared to "Noise : <0.1mV" for the Kubota, both falling considerably short of the standard you suggest. And it may be that neither is up to what is needed.
Fortunately the choice of regulator is highly modular in the power chain and thus particularly amenable to testing but if the Kubota can be rejected now, great. We will soon have participants with regulators ranging from genuine Borbely through Chorbely to Kubota. The SuperTeddyReg has repeatedly been suggested but I have no comparative noise measurement for it. Perhaps we should consider a pair of beefy 24vDC batteries as well. Anybody know the noise levels for 1)genuine Borbely, 2)SuperTeddyReg, and 3)truck or solar batteries?
Apply the gain of a +28dB power amp and you have <1250uVac of noise at the speaker.
For a 50w Amplifier that is just -84dB ref. max power.
That would be atrocious in my book.
But we're not talking about power amps in this thread so you said is not relevant.
You're forgetting that the line stage circuit has some kind of power supply rejection ratio which would reduce the effects of power supply noise. A good op amp has a PSRR of 100 dB! So, that "noise" from the power supply is comprised of ripple and a small amount of random hiss. This will be attenuated by 100dB by the line stage, if it's a good one.
Last edited:
I believe that even these high performance batteries have higher source impedance than a good power supply regulator. Even large capacitors with low ESR have more source impedance than a good regulator.
- Status
- Not open for further replies.