Thanks for the reply Hesener. I have the smds in place and the output stage is definitely not responsible for the oscillations.
It probably is the Hfe. Of course this useful advice could have come at a slightly earlier stage. Joachim recommended the highest Hfe devices to be used at the input...
At this stage i will probably do something more productive with my life and leave some other beta-testers to try their luck. This is a breadboarding exercise and not something meant to be done on a fully populated board, right?
It probably is the Hfe. Of course this useful advice could have come at a slightly earlier stage. Joachim recommended the highest Hfe devices to be used at the input...
At this stage i will probably do something more productive with my life and leave some other beta-testers to try their luck. This is a breadboarding exercise and not something meant to be done on a fully populated board, right?
That's what we expect from you, and how the shunt actually behaves in reality ;-) I am still considering doing a separate PCB for it (if you don't mind of course) because it can be used in many different applications, due to its very good quality
Was going to keep that hush hush and do same but now I am looking forward to your board especialy as quality of paradise boards is stunning (I think I can say that with no danger of ridicule now that a few started to play with it.)
And yess PSU is prety good to (bit like Londoner tha is allright if you catch my drift)
If I may? can power devices be on the edge as then one has option to use outside case sinks?
Put my name down for 4 please
Hi Alfred
Maybe you could post a schematic with the correct placement for those ceramic caps.....
Originally Posted by analog_sa
Fixed the issue with the insufficient voltage across the output 2SK170. Indeed Joachim was correct - top FET needed replacement.
Fixed the issue with the insufficient voltage across the output 2SK170. Indeed Joachim was correct - top FET needed replacement.
Hi,
from a schematic point of view its quite easy - they go from VPLUS (or VMINUS) to GND. The secret is in the placement - as close as possible to the emitter resistors of the current mirrors. Take a look at the bottom side of your PCBs and you 'll see what I mean
They don't work at IDSS. Probably a suitable Gm is needed. Did mine by trial and error.
Not sure but seems that varying the 220ohm resistor in the CCS may be another way of fixing this. In any case the CCS needs to output less than the K170 Idss.
It is becoming clear that the assembly guide can do with some added detail. While it's possible that the Paradise would work properly after just stuffing the board, it seems quite unlikely.
A warning that the circuit is prone to oscillations and some clear advice on how to avoid them per example seems absolutely essential.
The original advice on beta selection seems to also have been superseded now.
It is inconceivable that all builders will be reading through hundreds of posts from various threads in order to obtain such basic guidelines.
Would Erno Borberly's jfet preamp article , which includes a cascaded jfet input be helpful in explaining the needs of the top jfet.
http://ken-gilbert.com/images/pdf/erno_borbely_on_jfets.pdf
http://www.ltwiki.org/images/b/b1/Erno_Borbely_fet_articel_2.pdf
First article is the one that speaks to the cascode situation.
http://ken-gilbert.com/images/pdf/erno_borbely_on_jfets.pdf
http://www.ltwiki.org/images/b/b1/Erno_Borbely_fet_articel_2.pdf
First article is the one that speaks to the cascode situation.
Last edited:
Skip loot if you like butread bold part sorry for ranting a whole page
I agree with you about more detail and guidelines for oscillation on the guide IMO this need to be updated with advice considering the need that a wide skill set is out there with 200 + builds taking place.
I don't agree on, as you point out, on the unlikely.
As a beta tester I built Paradise R2, I know of my personal correspondence with Joachim and Alfred, from this personal experience you can be assured that final (for now) R3 version would have never be relesed if any problem became evident in my build, build that was carried out in a rush without particular care in regards to matching, I never measured any of the J fets for example and purposely mixed up Hfe to the 10% tolerance as specified, I am also on R2 using very hi Hfe trough out the board as the % volt test circuit revealed once I started using it.
Only problem I had with this built was that I placed one of the electrolytic the wrong way around, this is easy to do so I suggest any one building paradise take a break before stuffing the boards, You been warned
As for the apparent problems arising from the component kits, I did purchase one by the way, I cannot comment, as I don’t know what happened there,
Something is gone wrong certainly but I trust that steps are taken as we write to rectify the situation.
Would I have volunteered for such a task knowing of the 200 + builds?
NO WAY MATEY not for 200, maybe 20
500 transistors for 200 builds equal 100,000 pieces
Whit this large an order I would have gone directs to Manufacturer but is easy now to say.
Spread between Hfe in different goodies bag is normal, nothing that can be done about it, gett in touch with mate near you and swap. and make a few new friends in the process (I am doing that already) and no don't worry nothing wrong with me
The fact that BCXX-25 S Hfe are there is not acceptable
Where the blame need to fall need only to be established so we can take constructive actions to solve the problem.
This is DIY, problem solving is all part of the fun, if not for this and the satisfaction to have something one made that work beautifully as My R2 does, maybe one should order ready-made gear from the Shop whit all that it entails
Skip loot if you like butread bold part sorry for ranting a whole page
Goodness, you seem to enjoy mangling this language
Do you by any chance remember the beta range you used in the front end and mirrors? I take it your build shows no oscillations on the scope.
about 500 as that is the range of one I have direct from Fairchilds
the PNP may be higher as I mentioned no special care taken to keep even with people that actualy listen to music as main free time occ.
PS Skip the humor and instead reconsider your harsh comment as sems I am wasting time for no rewards or you end up as 5th member on my people to ignore list
the PNP may be higher as I mentioned no special care taken to keep even with people that actualy listen to music as main free time occ.
PS Skip the humor and instead reconsider your harsh comment as sems I am wasting time for no rewards or you end up as 5th member on my people to ignore list
Last edited:
yes, and we all learn with this build - its great you are sharing your experience!!
In fact, we (the MPP team) went through TWO betabuild sessions (R1 and R2) and thought that the oscillation problem was gone with the R2 layout + reduced number of mirrors. While I thought that the high HFE could be a indication of the problem, I did make tons of experiments with R1, sadly none of them worked. There is only one set of R1 boards (a.k.a., mine) that worked without oscillation. Joachim had the boards over at his house so he can confirm.
RCruz kindly provided his boards to me, and I did "massacre" them a lot, trying all kinds of things (this is where the famous "ground spider" came from). I did swap the input stage transistors with a new set of transistors that Le'Flu kindly provided, that had HFE in the range of 400, but the oscillation remained.
At that point my conclusion was that all transistors (input stage and mirrors) should be at 400 or lower, to avoid the problem. Then came R2 (which essentially is identical to R3), and in went my transistors with HFE from 400 to 450. No oscillations. We all thought the problem was solved.
Now, considering the new inputs, in my opinion ( and that is really just my opinion), it is best to not use too high HFE in the build. If you can keep it below 450, thats best. Especially in the input stage. Tight matching really is more important than anything else. I know that is not what some of you builders may want to here, especially with 400+ transistors on your desks that all measure in excess of 450 (I had some exceptional transistors that had 600 and 650....)
Let's see how you get on, it may be a special problem with your build. You might want to do me a favour and remove Q22 and Q44, and replace with a wire between B and E (the two traces that are running on the top side of the PCB), just to see if that makes a difference.
Yes, Bksabath was one of our guinea pigs and he did well
You are absolutely right, I did implement inputs from 5 or 6 betabuilders including myself, we did a lot of testing. Joachim was VERY insistent on only releasing a board that has its infancy problems removed, and is good to go for the wide skillset out there....
same here..... too late now....
500 transistors for 200 builds equal 100,000 pieces
Whit this large an order I would have gone directs to Manufacturer but is easy now to say.
Spread between Hfe in different goodies bag is normal, nothing that can be done about it, gett in touch with mate near you and swap. and make a few new friends in the process (I am doing that already) and no don't worry nothing wrong with me
The fact that BCXX-25 S Hfe are there is not acceptable
Where the blame need to fall need only to be established so we can take constructive actions to solve the problem.
This is DIY, problem solving is all part of the fun, if not for this and the satisfaction to have something one made that work beautifully as My R2 does, maybe one should order ready-made gear from the Shop whit all that it entails
You're right.
For a "commodity" like this (BC337/327) that has been around for a long time, I am hard pressed to think of a manufacturer that would actually agree to produce a larger volume of HFE-binned devices. This would be a very expensive exercise, and we should have started the planning process with this (could easily eat up one year or so, most of the time convincing people). The manufacturers would simply lean back and say "just buy 20x the quantity and check for yourself", which is what we are doing... A semiconductor test house might be an option, but again, expensive.
Thanks Hesener, will do this shortly.
As regards to the input stage PS decoupling it seems the onboard films are quite some distance from the actual devices. Do you think some decoupling right at the junction point of the 33ohms will be beneficial? I will probably try it anyway.
I used 540hfe for the input, 530 for the 1st mirror, 550 for second mirror, 560 for the central darlington and 580 for the shunts.
Never had any oscilation issues... so after all how can you state this is prone to oscilate ?
My only issue was on one of the output buffers where I obvioulsly was not carefull with choosing the correct fets.
To my knowledge all beta builds worked right from the start and that is why it was agreed to start a GB.
Decoupling fixed all oscillation issues. All that remains is a rapidly fluctuating offset and what i think is excessive noise floor with a DL103.
I doubt i will find a satisfactory solution to the offset. A quick listen without the servo seems to indicate the sound is preferable without it, so probably just a coupling cap will suffice for me.
I doubt i will find a satisfactory solution to the offset. A quick listen without the servo seems to indicate the sound is preferable without it, so probably just a coupling cap will suffice for me.
Good news !!!
You mean you decoupled VCASCH and VCASCL ?
What cap value / type did you use ?
Can you post a picture so others can learn how you did it ?