By grounding one input it would would be possible but many alternatives exist for converting single ended to balanced inputs so that the beauty of the JC-80 design might better be utilized.
I think you want around 0.1 uf for the bypass caps - .022 is a little small
The power supply - The diode bridges should not be connected to the ps ground and there should be a cap on each secondary, and not one between them as you show, these added caps should not be joined to each other or grounded. The ground only begins after the choke.
Use 1N4148 in series with 22 ohms for the diode protection from output to input on the 317/337 regulators as per Ben Duncan - lees capacitance than 1N4001 series so less hf noise gets past the diodes.
I prefer the op amp shunt regulator over the 317 type ( i actually used Borbely's discrete Jfet/ fet shunts, but that's hard to build without a good selection of jfets).
The power supply - The diode bridges should not be connected to the ps ground and there should be a cap on each secondary, and not one between them as you show, these added caps should not be joined to each other or grounded. The ground only begins after the choke.
Use 1N4148 in series with 22 ohms for the diode protection from output to input on the 317/337 regulators as per Ben Duncan - lees capacitance than 1N4001 series so less hf noise gets past the diodes.
I prefer the op amp shunt regulator over the 317 type ( i actually used Borbely's discrete Jfet/ fet shunts, but that's hard to build without a good selection of jfets).
Last edited:
Does it mean that Blowtorch -like power supply is not suitable for preamps where positive rail source more current to the circuit than the negative(eg. Pass Xono , 130 mv on pos. , 80 mv negative)?
CM choke saturation, perhaps, or results of two rec. bridges approach?
Mr. Wayne Coburn, Xono and Pearl phono designer, suggest using two bridge PS for preamps with unsymmetrical current drain in order to avoid possible noise induction.
As far as I remember, attached circuit is posted by a French guy who was trying to guess the original Mr.Curl PS design . Original Blowtorch P.S schematic has never been published since it is a trade secret.But, Mr. Curl detailed description is more than enaugh to get the whole picture. Just read carefully.
He does not use current source and IC referenced shunt regulator as Frenchman does. Instead, he uses 10 ohm series resistor and simple open loop shunt regulator, of unknown topology. Blowtorch PS seems to have LM317/LM337 in textbook topology, not in unusual fashion as in above circuit.
Thanks, Kamis, now things are beginning to make sense and I can stop looking for 'the' schematic. I must admit to some trepidation because the BOM that came with the JC-80 PCB left some components less than fully specified and I am insufficiently prepared to judge "how well the parts do the job assigned to them" in the PS. Having spent years as a software guy, where (re)wiring was merely an edit+compile and the results immediately testable, I find circuitry design intimidating.
The Hong Kong DIY community powers these boards with quite a variety of supplies. The customer service from the vendor I chose, i.e. the one offering the red PCB, is superb (and I have sampled both vendors before writing that) and easily worth the slightly higher price. I now believe both versions of the PCB to be identical.
Current Requirement
I am considering a single Kubota regulator capable of delivering 500ma. That seems quite a lot for 300R headphones but since the onboard regulators will dissipate some power as heat (dropping from 30v to 24v and then down to 15v), I hope that will be adequate. I do know of at least two cases where JC-80s are being driven by Borbely shunts with the same current capability but am not sure if that allows one or two JC-80s per shunt.
I am considering a single Kubota regulator capable of delivering 500ma. That seems quite a lot for 300R headphones but since the onboard regulators will dissipate some power as heat (dropping from 30v to 24v and then down to 15v), I hope that will be adequate. I do know of at least two cases where JC-80s are being driven by Borbely shunts with the same current capability but am not sure if that allows one or two JC-80s per shunt.
One regulator per channel
Adding a second Kubota will add less than 10% to total cost and there will then be no question of providing adequate current. Many of the parts are in my left overs and BJT pool.
The Asian parts are beginning to trickle in -- Caddocks arrived today. I expect the remainder to stretch out over the next week or two. During that quiet spell I will complete one Bal-Bal "wire" by OPC, for which everything except the PCB (mailed today) is already in-house.
Adding a second Kubota will add less than 10% to total cost and there will then be no question of providing adequate current. Many of the parts are in my left overs and BJT pool.
The Asian parts are beginning to trickle in -- Caddocks arrived today. I expect the remainder to stretch out over the next week or two. During that quiet spell I will complete one Bal-Bal "wire" by OPC, for which everything except the PCB (mailed today) is already in-house.
System Interconnect
There is enough space in the external power supply case for a second set of r-core, rectifiers, and caps. It would add considerably to the total cost and force me to use the pristine PS pcb (image in#36) so I will likely get one channel up completely first.
So far no source for the candidate PS pcb has turned up. There is one small error on the board that is easily correctable, no copyright, and enough flexibility to allow dual secondaries, plenty of rectifiers, caps, etc. The Kubota lacks that flexibility so, since the regulators will be mounted in close proximity to the JC-80 boards, I expect 3-4 feet of cable to the first ps. Are there special considerations for that interconnect? The Kubota pcb has only two conductors from rectifiers to regulator so the "ground" is established by the regulator.
If all this works as well as I hope, i.e. sounds better than the Bal-Bal wire, then I will 1:1 image the pristine PS pcb in case others get the urge to build everyman's JC-80.
There is enough space in the external power supply case for a second set of r-core, rectifiers, and caps. It would add considerably to the total cost and force me to use the pristine PS pcb (image in#36) so I will likely get one channel up completely first.
So far no source for the candidate PS pcb has turned up. There is one small error on the board that is easily correctable, no copyright, and enough flexibility to allow dual secondaries, plenty of rectifiers, caps, etc. The Kubota lacks that flexibility so, since the regulators will be mounted in close proximity to the JC-80 boards, I expect 3-4 feet of cable to the first ps. Are there special considerations for that interconnect? The Kubota pcb has only two conductors from rectifiers to regulator so the "ground" is established by the regulator.
If all this works as well as I hope, i.e. sounds better than the Bal-Bal wire, then I will 1:1 image the pristine PS pcb in case others get the urge to build everyman's JC-80.
FIRST, make sure that the local cap followers in the power supply are separate. SECOND, make sure that any pre regulators are separate, Third, if practical, make separate rectifiers and power supply caps separate, BUT this is NOT mandatory. However, the effect of not doing so can be measured at very low frequencies. 'Stereo Sound' in Japan did just that on the original JC-80 in their review.
Last edited:
Some clarity?
My statement above: "The Kubota pcb has only two conductors from rectifiers to regulator so the "ground" is established by the regulator." is incorrect. The schematic that accompanies the product only shows it that way, leaving out the transformer center tap ground, but the pcb is correct.
I have attached an overview schematic for reference. The act of drawing it allowed me to put together a number of earlier comments, clarify my own thinking, and (at last) see why extending the symmetry back into the roots is important.
The external block incorporates earlier suggestions but raises two questions. 1) will thermistors add noise and 2) should two CM chokes be applied before joining the upper and lower rectifiers pathways to ground?
Bouncing all the way to the right, capacitors C2 & C3 (all ids are relative to the silk screen of pcb) were discussed earlier and ticknpop wrote:
The caps I have ordered for C2 & C3 are 0.47uf radial polystryene that I plan to place where the parallel electrolytics would have been. I assume C12 & C13 are the electrolytics referenced in the final sentence of your quote, ticknpop.
I had tried to formulate questions about the appropriate cabling between the enclosures. It now seems to me that two shielded twisted pairs should extend back, one from each regulator, to the external rectifier pcb(s). Given that the + and - conductors would carry equal but opposite currents, crosstalk should cancel. If a single pair was used this cancellation would be imperfect.
My statement above: "The Kubota pcb has only two conductors from rectifiers to regulator so the "ground" is established by the regulator." is incorrect. The schematic that accompanies the product only shows it that way, leaving out the transformer center tap ground, but the pcb is correct.
I have attached an overview schematic for reference. The act of drawing it allowed me to put together a number of earlier comments, clarify my own thinking, and (at last) see why extending the symmetry back into the roots is important.
The external block incorporates earlier suggestions but raises two questions. 1) will thermistors add noise and 2) should two CM chokes be applied before joining the upper and lower rectifiers pathways to ground?
Bouncing all the way to the right, capacitors C2 & C3 (all ids are relative to the silk screen of pcb) were discussed earlier and ticknpop wrote:
The caps I have ordered for C2 & C3 are 0.47uf radial polystryene that I plan to place where the parallel electrolytics would have been. I assume C12 & C13 are the electrolytics referenced in the final sentence of your quote, ticknpop.
I had tried to formulate questions about the appropriate cabling between the enclosures. It now seems to me that two shielded twisted pairs should extend back, one from each regulator, to the external rectifier pcb(s). Given that the + and - conductors would carry equal but opposite currents, crosstalk should cancel. If a single pair was used this cancellation would be imperfect.
That's the way the pcb was designed. It will allow dual secondaries but it joins and grounds the bridges before the caps. John may have suggested keeping the two separated through the caps when he wrote "FIRST, make sure that the local cap followers in the power supply are separate." There are ten sets of cap pads but one side of each is the ground plane :-(
I have looked at several split rail schematics and it is not uncommon for the zero potential to be grounded. It does look unusual but my goal in drawing was to understand the interconnect. I have had quite a time blowing fuses due to inrush so I did want to include thermistors and the CM choke and ask if they were placed in the correct order.
The Kubota does not allow for dual secondaries and its flexibility for rectifiers and caps is similarly constrained. I hope the external pcb will prove OK for this application.
The Kubota does not allow for dual secondaries and its flexibility for rectifiers and caps is similarly constrained. I hope the external pcb will prove OK for this application.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.