Cheers everyone!
I finished the first stereo set using the last batch of boards. They worked right away from the first power up (after correcting the bug signalled by Emil) and the performance is outstanding. I have to congratulate the whole team for the project, and especially Alfred (hesener) for the PCB artwork and documentation, because his layout helped me to implement this build as I wanted. I also want to thank every builder that shared the details/issues and contributed to the Paradise builders and all related threads.
On top of the existing documented features, I had a few additional goals: to have the input transistors thermally coupled, to have them cooled, to have the mirror transistors coupled and to have them cooled. I managed to get my targets accomplished, although the cooling is just passive (as a side discussion, probably the next set will have the input stage cooled by Peltier elements). I strongly believe that one should aim for stable DC conditions, so that the output servo must have very little to work!
I will post later a few details of the build, but I beg for mercywhile evaluating my build, as I am not a skilled mechanical engineer, and nothing was planned and implemented meticulously. Also, to have this accomplished, I rotated all the transistors and mounted them down of the board. This is a very tough decision, as you are not allowed to make mistakes. Of course, everything is "undoable", but requires a lot of work to correct mistakes. For builders without patience, or people that are in a big hurry to finish, I suggest to use the well documented guide, and skip my approach. However, I guess it is worth to have a different approach and think before installing a component, as it spares lots of time and emotions in the debugging process.
I have built an aluminium frame using L profile 15x15x2mm. I have drilled transistors holes using 4,5mm drill in a 40x4mm aluminium plate. That was a little tight and I have adjusted every hole with a round file; maybe 4,8mm holes would have been better. At least Diotec and Fairchild transistors used have an incomplete cilinder case with a diameter of 5 mm. Each transistor was forced inserted in the hole using a small desk vice. Having pieces of aluminium with 24 terminals each, it was not a nice task to fix them on hte PCB and the frame, but it is doable. I guess two days I spent only aligning these mechanical "small" details.
I have tested first the DC conditions. After few minutes they were stabilized very well. The offset at the input was 0.0V with the potentiometer set to the middle of range; using 100Ohm input load and 10Ohm input load made really no difference. In the test points, differences between positive and negative branches of a channel were in the milivolt range, and a big more for differences between channels in the same test points. Although I used input transistors with an hFE of 568 (measured with Atlas in the same day, same temperature, and under a blowing fan) I have no significant offset at the output. I measure around 0,2mV, fluctuating within +/- 3mV range. AC at the output, measured with a DSO having a 30MHz range was a clear line while the input had only the load resistance of 100Ohm. No oscillations and no problems at all (OK, I have a pocket scope, with only 50mV/div sensitivity, but anyway, able to detect a problem within 30MHz range).
After I left it running for one day, and being sure that everything is OK and safe to further testing, I have disconnected the preamp from the test room and brought it to the listening point. It took me some half an hour to make all the connections, after which I left a Status Quo record to play.
Terrible sound!!! With a cold shunt supply, for me it sounded like an old transistor portable radio. Imaging was really good, but the sound was crispy and lifeless. You can figure out how I felt, after so many weeks of work. Anyway, desperation is not something for me, and I listened further. For every minute passing by, the sound was improving a lot. After the first side of the disc it was comparable to the one from the integrated head amp in my Yamaha CX1000. Patiently, and avoiding any drop of alcohol - thus eliminating the subjectiveness from my judgement - I continued to listen, avoiding the first impression impact.
After around another hour, the sound was magnificent. Seismic bass and silky trebble, and a very good localization. I am not a good person to analyse an audio product, but the difference between startup and one hour running time is that HUGE! According to my quite good memory, this is the best phono preamplifier I have ever heard. The existing one in CX 1000 is also a good one. It uses two pairs of 2SB737/2SD786, but yeah... global feedback, input coupling cap (100uF electrolytic)... Paradise is clearly more silent and it's a big sounding difference between them, soundwise. I avoid to compare it to other preamps I listened in the past, as probably I cannot be objective.
As a source I use a Denon DP57 in the original state, equipped with Goldring Elite, 8Ohm DC resistance, and an excellent sound for the music I like to listen (mostly METAL, but also other genres guitar based).
Sorry for the long post; I just want to confirm what Joachim said: if you think every move in the build, it can/will run from the first power up. I will add later my construction details, with some pictures, as now I still need to listen. It's really addictive
Nice review. Totally agree wrt warming up . The sound is so much better once it had about an hour or so turning on. I would not be brave to leave it on permanently.
Well, the paradise PCBs need 26V minimum at their inputs so the onboard regulators can work properly. What you connect there is not really important, as the onboard regulators will do their job. I looked at the link you added and it should be no problem, provided you have the proper input voltage.
My build details
I come back with some pictures of my build, as promissed.
The frame that supports the board, built from aluminium L profile 15x15x2mm:
Transistors are forcefully inserted in aluminium plates 4mm thick and 40mm wide:
The holes should be drilled with 4,8mm drill bits for a firm contact and an insert without victims among the transistors. I used a 4,5mm drill bit and used a file to enlarge the holes. During the whole process I destroyed 2 transistors
The plates with transistors are connected to the frame using threaded holes in the plates and 3mm screws:
Plate is seating quite tightly fit in the frame, screws mounted to hold also the plates tightly and the transistors can be soldered on the board, after a long process to fit all the terminals in the holes:
The boards are fitted in the box using double sided adhesive tape. To be able to mount and unmount the boards, the bottom of the box has 8mm holes drilled in the position of the frame holes. I did not take too many pictures of the process, only the PCB/frame unit sticked on the plate. i know, it is not something professional, but the tape provides isolation from the air movements and the frame/plates provides cooling for the transistors. These are all placed under the shunt heatsink level, so the heatsink does not heat the active components placed under the PCB in the amplifier stage:
Also, the double sided tape provides mechanical damping to vibrations, as the frame is not connected wit screws to the bottom plate of the enclosure.
Paradise before closing the lid, after all DC conditions were tested:
I come back with some pictures of my build, as promissed.
The frame that supports the board, built from aluminium L profile 15x15x2mm:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Transistors are forcefully inserted in aluminium plates 4mm thick and 40mm wide:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The holes should be drilled with 4,8mm drill bits for a firm contact and an insert without victims among the transistors. I used a 4,5mm drill bit and used a file to enlarge the holes. During the whole process I destroyed 2 transistors
The plates with transistors are connected to the frame using threaded holes in the plates and 3mm screws:
An externally hosted image should be here but it was not working when we last tested it.
Plate is seating quite tightly fit in the frame, screws mounted to hold also the plates tightly and the transistors can be soldered on the board, after a long process to fit all the terminals in the holes:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The boards are fitted in the box using double sided adhesive tape. To be able to mount and unmount the boards, the bottom of the box has 8mm holes drilled in the position of the frame holes. I did not take too many pictures of the process, only the PCB/frame unit sticked on the plate. i know, it is not something professional, but the tape provides isolation from the air movements and the frame/plates provides cooling for the transistors. These are all placed under the shunt heatsink level, so the heatsink does not heat the active components placed under the PCB in the amplifier stage:
An externally hosted image should be here but it was not working when we last tested it.
Also, the double sided tape provides mechanical damping to vibrations, as the frame is not connected wit screws to the bottom plate of the enclosure.
Paradise before closing the lid, after all DC conditions were tested:
An externally hosted image should be here but it was not working when we last tested it.
You're welcome! I thought also of a better solution where you can remove the PCB out of the box with transistors only, so that you can service the PCB much better in case you need, or if you want to fine tune the RIAA components. However, I could not find any practical idea and I abandoned it in favor of the presented solution. If you like fine tuning, maybe an external RIAA should be implemented at the provided connector.
Anyway, provided the extremely good offset result, I guess it is worth implementing this.
Anyway, provided the extremely good offset result, I guess it is worth implementing this.
Me too, but the linkage posts FdW puts up help a lot. Post 2213.
I didn't understand the board mod pic until I looked at my boards. If it looks like this you need to cut around the pad.
All you need to do is skim it gently with a decent 6mm HSS drill bit, just enough to break the contact between the via and ground plain.
An externally hosted image should be here but it was not working when we last tested it.
Thanks Alfred I will try that and see if it makes any difference, it's an easy swap.
i'm going to use sk170's in my next build is there any preferred grade or spec for these parts?
Thanks Joachim, and many thanks for sharing this high end design with the community.
I have 8 pairs of low Rbb' transistors (B737/D786) with an Hfe of around 315 (+/-3) as measured with Atlas at 21C room temperature, and I would like to use them in the next build set. Together with Peltier cooling, I would also reduce the Emitter resistors. What do you recommend to use as the lowest safe value?
I have 8 pairs of low Rbb' transistors (B737/D786) with an Hfe of around 315 (+/-3) as measured with Atlas at 21C room temperature, and I would like to use them in the next build set. Together with Peltier cooling, I would also reduce the Emitter resistors. What do you recommend to use as the lowest safe value?
Hey guys how important is correcting this resistor if boards are mounted and sounding excellent??