I for one, as chronic capacitance multiplier (ab)user, am courious of this, already using couple of Didden superregs for DAC builds to good effect.
Or maybe just looking for excuses of stuff to build
What is design requirement for the pass mosfet? Low RDSon and what else? Probably suitable replacement could be found..
Or maybe just looking for excuses of stuff to build
What is design requirement for the pass mosfet? Low RDSon and what else? Probably suitable replacement could be found..
It’s not that only MOSFETs are out of stock. At any moment, several parts are not available. However, over period of several months most parts appear in stock briefly, just not at the same moment, so there is a chance to gather them.
Different replacement parts can’t be recommended without thorough design reevaluation. Every semiconductor’s different parameter can have a large negative effect. For instance, replacing KSC3503 with BD139 results with worse transient behavior below 1V voltage drop, while with KSC it is very good, even at mere 0.15V. For MOSFETs everything counts: SOA, RDSon, max. pulse current … and especially Ciss. There are suitable replacement parts, but even those don’t have stable stock.
Different replacement parts can’t be recommended without thorough design reevaluation. Every semiconductor’s different parameter can have a large negative effect. For instance, replacing KSC3503 with BD139 results with worse transient behavior below 1V voltage drop, while with KSC it is very good, even at mere 0.15V. For MOSFETs everything counts: SOA, RDSon, max. pulse current … and especially Ciss. There are suitable replacement parts, but even those don’t have stable stock.
Hm! Was a bit afraid it wouldn't be so simple.
Guess we will have to wait and see when stock frees up.
Or maybe i make ltspice model and play a bit, but i suspect this would be fruitless.
This is real nice work you did, it looks like a lot of evaluation went into every part, and i wonder how many iterations..
Guess we will have to wait and see when stock frees up.
Or maybe i make ltspice model and play a bit, but i suspect this would be fruitless.
This is real nice work you did, it looks like a lot of evaluation went into every part, and i wonder how many iterations..
Only breadboard test can confirm that replacement part is suitable. LTSpice is to confirm that part “could” work as expected. Circuit stability evaluation with Tian probe is far from real life results.
LTSpice example: calculation of low frequency PSRR was at 170 – 180 dB levels, if C6 was 470 uF. Wow!
After 3 PCB revisions and several measuring equipment verifications, I got it that it was an optimistic calculation with not so complete model.
So yes, it was immense fun all along, before arriving at final version.
I’m sure a small company wouldn’t be ashamed of all verifications and tests I ’ve done.
Of course, i agree. LTspice is not final say, just 'cutting short design and evaluation process significantly' tool. And make you feel better about what you're building with optimistic result. Especially when you go check THD but half of models don't have half of things modelled
Yes sounds like that's about the scope of work you invested here, honestly i do not see many "super regs" for power amps on here so surprising there is not more interest. You even provide everything.So yes, it was immense fun all along, before arriving at final version.
I’m sure a small company wouldn’t be ashamed of all verifications and tests I ’ve done.
It is just a matter of changing voltage divider R10/R11 ratio. Changing R10 to 5100 Ω would enable output up to 80 V. As mentioned in the build guide, at high output voltages, dissipation on some resistors would be unnecessary high. It would be better to change R10 to 10 K , R11 to 6K8 and R5 to 33 K.
Ha! Yes, that's exactly what I was thinking about, but I'm not sure it will work out, given how difficult it is to source the parts.
I'm also not 100% certain how much benefit it will be (at least to me). The amp is designed with beyond-stellar ripple rejection and noise performance built-in. The figures they achieved were measured using a simple 10,000uF per-rail C-supply. A beefier C-supply would provide plenty of current and excellent transient response. So the main benefit from implementing this would be heard at the extremes, no?
Last edited:
Regulated power supply will not change measured amplifier distortion. There is a reason why dual-mono PS is touted as a “way to go”, if one wants best soundstage, as many amplifier designs have mediocre or low PSRR. Regulated PS allows single PS usage for both channels and to still have perfect channel isolation.
On an amplifier design with very high PSRR, like Wolverine, improvement is doubtful.
There are other effects on dynamic rail loading like harmonics (not from the rectifier but from the rail load changes), as explained by people that know much more than I do:
https://www.diyaudio.com/community/...ith-ic-voltage-regulators.359652/post-6329354
Whether those rail harmonics could affect sound is again a matter of amplifiers’ PSRR and listener’s perception. There are members here that can hear a change if power switch or relay is involved, different rectifier diodes, PS capacitors etc. I don’t discard their claims, but I do keep a healthy skepticism.
So, YMMV.
On an amplifier design with very high PSRR, like Wolverine, improvement is doubtful.
There are other effects on dynamic rail loading like harmonics (not from the rectifier but from the rail load changes), as explained by people that know much more than I do:
https://www.diyaudio.com/community/...ith-ic-voltage-regulators.359652/post-6329354
Whether those rail harmonics could affect sound is again a matter of amplifiers’ PSRR and listener’s perception. There are members here that can hear a change if power switch or relay is involved, different rectifier diodes, PS capacitors etc. I don’t discard their claims, but I do keep a healthy skepticism.
So, YMMV.
Thanks! As I thought. Better off sticking to a large capacitor bank for the Wolverine. But I can use your awesome design elsewhere. I'm tempted to put in an order for the parts so I get them whenever they're available (a year or so, at this rate).Regulated power supply will not change measured amplifier distortion. There is a reason why dual-mono PS is touted as a “way to go”, if one wants best soundstage, as many amplifier designs have mediocre or low PSRR. Regulated PS allows single PS usage for both channels and to still have perfect channel isolation.
On an amplifier design with very high PSRR, like Wolverine, improvement is doubtful.
There are other effects on dynamic rail loading like harmonics (not from the rectifier but from the rail load changes), as explained by people that know much more than I do:
https://www.diyaudio.com/community/...ith-ic-voltage-regulators.359652/post-6329354
Whether those rail harmonics could affect sound is again a matter of amplifiers’ PSRR and listener’s perception. There are members here that can hear a change if power switch or relay is involved, different rectifier diodes, PS capacitors etc. I don’t discard their claims, but I do keep a healthy skepticism.
So, YMMV.
Trying to put together parts for two positive R21 boards, and SO close. I can account for nearly everything from Mouser and Digikey. ZTX958 and VOM1271T I can source from eBay. That leaves SUP70101EL-GE3, which I cannot find anywhere. Anybody know where two pieces may be hiding? Or if a substitute has been identified?
No suitable replacement was identified. Situation with parts availability is terrible and will remain so. Imagine how many electronic parts are required for modern weapons.
There are two solutions:
There are two solutions:
- Mouser has over 20000 on order. Buy now and wait. It should be before end of year.
- I can send you 2 pcs. used during development, but in perfect conditions, just soldering residue visible. That leaves me one spare SUP70101, which is fine with me.
That’s very generous Tombo, thanks for the offer!
Let me ask you your opinion of my use case. I have single rail CRC for my 77B Singing Bush mono blocks. Each channel currently has 47,000uf > .15R > 47,000uf. While I currently have no complaints with noise or performance, I think the parts deserve my very best efforts. Do you imagine audible performance improvements could be accomplished using your module in this application?
Let me ask you your opinion of my use case. I have single rail CRC for my 77B Singing Bush mono blocks. Each channel currently has 47,000uf > .15R > 47,000uf. While I currently have no complaints with noise or performance, I think the parts deserve my very best efforts. Do you imagine audible performance improvements could be accomplished using your module in this application?
Hard question as I don’t have circuit available for the measurement. From the theoretical standpoint, you are good as is. Singing bush has CCS and draws almost constant current from the supply rail, even with large output signal. Hah, Mark Johnson made an explanation here:
https://www.diyaudio.com/community/...citance-multiplier-output.389938/post-7114591
From the nitpicking standpoint, CCS is not ideal and its impedance falls toward high frequency, so there is some small leakage or influence from the PS at the higher frequencies. Can’t estimate how much as I don’t have Spice model for the IXFN140N20P. But, with 47.000 uF, that shouldn’t be anything audible. We are likely speaking of micro volts up to single mV.
https://www.diyaudio.com/community/...citance-multiplier-output.389938/post-7114591
From the nitpicking standpoint, CCS is not ideal and its impedance falls toward high frequency, so there is some small leakage or influence from the PS at the higher frequencies. Can’t estimate how much as I don’t have Spice model for the IXFN140N20P. But, with 47.000 uF, that shouldn’t be anything audible. We are likely speaking of micro volts up to single mV.
Ok, that makes sense. I suspected it wasn't an ideal candidate, but wanted to get a second opinion. Thanks! I think I will go ahead and order parts and implement in another project. Its not a big rush, so I'm ok waiting for the restock of SUP70101EL-GE3. Really do appreciate your generous offer though.
Thanks Tombo!
Thanks Tombo!