Working on another couple Paradise boards, one is working fine but the other is stuck at 15V at the output.....
Most interested in your findings... Had a similar issue that took me a while to correct... mybe you find a better way 🙂
Boards are working fine now, and the error was quite interesting.
A lesser-known fact is that bipolar transistors work both ways (collector and emitter reversed). the current gain is low, but if the circuit does not depend so much on this, at first glance it looks fine.
In this case, Q15 and Q16 were inserted the wrong way. They are the "bottom" transistors of the current mirror used in the current source for the input stage, and since this is a stacked current source , and the transistors were still doing something, it was not immediately obvious. Instead, there was some offset, which I thought was caused by mismatch in the transistors. But it was a mismatch between the upper and lower current mirror, caused by the reverse mounting. The servo could not correct this amount of mismatch - an indication how weak it actually is. Once the transistors were mounted correctly, everything worked fine.
Many thanks Ricardo for the numbered drawing, I used it extensively in this case!
Frans,
I'm doing a bit of catching up on this thread. Regarding your post "#1772 NJF selection for the PSU" - what about J109? 😀
Regards
Sven
I'm doing a bit of catching up on this thread. Regarding your post "#1772 NJF selection for the PSU" - what about J109? 😀
Regards
Sven
Not possible, the J109 needs an Rgs of more then 5K for 1mA, thus dropping more than 5Volts (no more than 2Volts (or less) is best).
I'm sorry - I'm one of these "color by numbers" guys. I don't really understand what happens and I need simple instructions 😀.
You write the J310 is not ideal. I assume Rgs is R108 (3k3R) in my board? So what I have today is a J109 (came with the GB) and 3k3R. Please be certain you do not mistake the J109 for something else, we caused some confusion earlier in this thread because of this 😉. Anyways, I take it from your answer that this is "not possible".
So you suggest SK170. I will have to replace the 3K3R with something around 360R? Does it matter if SK170 BL or GR?
Thanks for you patience.
Sven
You write the J310 is not ideal. I assume Rgs is R108 (3k3R) in my board? So what I have today is a J109 (came with the GB) and 3k3R. Please be certain you do not mistake the J109 for something else, we caused some confusion earlier in this thread because of this 😉. Anyways, I take it from your answer that this is "not possible".
So you suggest SK170. I will have to replace the 3K3R with something around 360R? Does it matter if SK170 BL or GR?
Thanks for you patience.
Sven
Last edited:
IMO the best choice is the J113 (using a 1K Rgs resistor). But all resistors needing a 2K or less Rgs are good. The J109 is currently the worst one simulated 🙂
Attachments
And, trying to un-confuse this a bit more, to be more clear. The 'impossible' means that it is impossible to have a 1mA current source (in this application) using the J109 and the 1mA was the design parameter for this current source.
On the other hand, the power supply will happily function using the J109 and 3k3, but; the supplied current will be between 600 and 750mA (not the design goal) and it will not be a CCS (not the design goal). Especially the last (it not being a CCS) hurts the performance of the PSU (but slightly).
These are the most important/relevant messages regarding PSU and Pre-Reg.
- Mpp #4375 PSU-pre-regulator.
- Mpp #4610 V2 Schemas UPS and Amp.
- Mpp #6506 CCS fine-tuning.
- Mpp #6462 PSU with my notes.
- Mpp #6809 Do not try to feed the CVS of the CCS of the shunt with a regulated power supply.
- Mpp #7127 More about transformer selection.
- Mpp #7635 Single and dual transformer wiring (also corrected drawing) (See also #7733).
- Mpp #7733 Grounding (PSU, RIAA and Player).
- Mpp #7810 35V PSU for the CroMagnon.
- Mpp #7830 The PSU explained.
- Mpp #8386 My resistor recommendation’s.
- My Paradise #56 PSU Noise.
- PradiseBuilders #13 Schema's and assembly guide.
- PradiseBuilders #155 Vdc input voltage minimum simulated.
- PradiseBuilders #970 PSU oscillates; posible solutions?
- PradiseBuilders #983 Compensating the PSU for oscillations.
- PradiseBuilders #1173 Fixing oscillations (up to #1182).
- PradiseBuilders #1322 Fuse 🙂.
- PradiseBuilders #1387 Updated Paradise R3 assembly guide (Also German).
- PradiseBuilders #1403 Capacitors bad vs good.
- PradiseBuilders #1593 PSU Power Transistor selection.
- PradiseBuilders #1644 PSU Output Impedance.
- PradiseBuilders #1680 PSU Power Supply Suppression Ratio(PSSR).
- PradiseBuilders #1772 NJF selection for the PSU.
- PradiseBuilders #1887 Empirical NJF Rgs determination.
- PradiseBuilders #2127 More about the PSU fets (J113's etc).
Some time ago it got an R1 board back from Hesener that still had some bugs.
AS Audio helped me to solve it.
One board had two transistors reversed and the oscillation was stopped by using base stoppers on the input of the bootstrapped diamond buffer.
Will be interesting to compare the R1 to the R3 now.
AS Audio helped me to solve it.
One board had two transistors reversed and the oscillation was stopped by using base stoppers on the input of the bootstrapped diamond buffer.
Will be interesting to compare the R1 to the R3 now.
Thanks for the hint - too bad my R1 boards never oscillated.... Yes, it would be interesting to compare sonic signatures. Also, I was thinking about putting the "Calvin buffer" into my paradise R3.
With so many Paradise stages out there, we should be able to come up with a "Paradise Fest" 😀
With so many Paradise stages out there, we should be able to come up with a "Paradise Fest" 😀
Alfred, when you implement the Calvin buffer please tell me how you like it.
I get really good results, especially more transparency and slam.
I get really good results, especially more transparency and slam.
I only have one channel of Paradise so far, but I have just shy of half of the "Calvin buffer" done [right hand schematic in #52]. I posted a picture in Calvin's buffer thread. It is scary looking, but small enough to fit in some empty space on the Paradise pcb. Now I just need to muster the courage to tackle the lower half and the output Zobel, a few more wires to interface with the world are also still missing.