When you speak about oscilation I guess you mean output offset wondering.... as long as it is on the mV range, I guess it might be due to amplifying noise from first stage.
this build must be enclosed to minimize out put wondering.... once it gets under 10mV there are no audible penalties.
I believe this is the price to pay for high gain dc coupling.
this build must be enclosed to minimize out put wondering.... once it gets under 10mV there are no audible penalties.
I believe this is the price to pay for high gain dc coupling.
None of us are talking merely dc drift here. It is a fast and wild fluctuation, cannot imagine enclosing to have any effect -it's in the hertz region.
Don't know about others but i also see something which may be oscillation in the MHz region. In any case the hiss level is unacceptable and my guess is that it does indeed oscillate.
If i get to play with the board again i would remove all the decoupling lytics from the input stage, then reintroduce them one by one and add decoupling across. Alas, more interesting projects at the moment.
My interest in the issue is mainly academic - i am indeed curious why some builds work only marginally. This can never become my main phono pre due to the input current disbalance. I don't know about others but i have 3 wands for my arm with different carts, ranging widely in impedance. If i put the wrong wand in the arm a 2k cart may prematurely go to Paradise. It is an accident waiting to happen and sooner or later it will. Such a device cannot be sold to a paying customer at any price. Being a diyer does not also make me completely mad.
As for the sound, it is indeed quite pleasant keeping in mind the modest component quality. It certainly does not obliterate my other phono stages soundwise.
Hard words for a non profit DIY project, many people spend there private time. I think you are old enough to decide build it up or not.
I left the DIY scene about for two years and saw the group buy late by random. Thus i read less about the concept only heard a predecessor in 2010 at a DIY festival. But never it was a secret for me about the risk of the design, for example oscillation.
We are DIY people and no plug and play customer, so we should solve the problems together.
To damage the cartridge it needs higher DC at the input. After trimming the input with a dummy, the offset on my board are zero with pluged cartridge.
Chris
Last edited:
No current should flow from preamp input into the cartridge coils. In other words, input bias current is not permitted. For this reason, JFETs at the input are much better than BJTs, like in SCP Vendetta Research.
True. The problem is that once you change the cartridge to another with a different coil resistance you will have to repeat the offset trimming.
Failing to do this is very unlikely to cause damage to the coils but it will quite possibly cause permanent magnetization - a real pain in the backside. It is even considered a bad idea to measure the coil resistance with an ohmmeter.
Huh? A current many times exceeding the ohmmeter's current will pass through the coils if you connect the wrong cartridge. You may still ask who told me to do so, but i and probably many others don't listen to a single cartridge.
Sorry if i am a teeny bit surprised with this design decision. And judging by the posts above not all builders are even aware of such a danger. Maybe two years down the line a third party decides to plug their turntable into this...
The Paradise has an input offset trim that reduces DC into the cartridge to minuscule levels. BJTs have a noise advantage under 1kHz because the 1/f corner is much lower then in J-Fets. The 2SJ74 is obsolete so we decided on a BJT input. We could have made a single ended Fet RIAA though or a differential input. Both options give higher noise though then parallel symmetric.
The maximum current that can flow into the cartridge, even without offset trim, is the idle divided by the hfe. The DC ohmic resistance of the cartridge plays a role too. This was discussed in detail before. That can neither destroy nor demagnetize the cartridge.
I have a quite valuable Lyra Titan i and it sounds like on the first day although i abused it with much higher current during my transimpedance input experiments.
We have a Paradise R2 playing since a long time in many systems and on several shows with different cartridges and people are reporting very good results. That Paradise did not have the offset trim. We included it later in the R3 version.
I have a quite valuable Lyra Titan i and it sounds like on the first day although i abused it with much higher current during my transimpedance input experiments.
We have a Paradise R2 playing since a long time in many systems and on several shows with different cartridges and people are reporting very good results. That Paradise did not have the offset trim. We included it later in the R3 version.
Censored comments about scare mongherers Censored
Built mine R2 with poor matching of transistors and ofset is steady at 0.0+-5 mV (0.05 )
with MC one use cartridge loading say 100 Homs and using formed capacitors helps again
So just get over it.
Ofsett is worth keeping in mind but do not panic over it.
Forming the caps is easy Just get a 317 to suply the nominal voltagge and live the lot there while the boards get populated
Built mine R2 with poor matching of transistors and ofset is steady at 0.0+-5 mV (0.05 )
with MC one use cartridge loading say 100 Homs and using formed capacitors helps again
So just get over it.
Ofsett is worth keeping in mind but do not panic over it.
Forming the caps is easy Just get a 317 to suply the nominal voltagge and live the lot there while the boards get populated
Last edited:
ok, thank you joachim. just to confirm: hfe around 500 in the input stage is desirable? any tradeoffs for high hfe there (hbarske stated in one post that hfs >500 provokes instability). in the current mirrors, what is the highest acceptable hfs (have none <480)??
i know, it all depends on a lot of other things, but is there something like a recomendation of optimal hfs-range, provided all pairs are tightly matched?
Ube is between .745-.752 for all devices npn and pnp. thats less than 1%, no need to bother with a closer match?
Last edited:
Fearmonger indeed. Not a chance in hell to kill a 2k cartridge with the paradise input bias current. The combination of matched devices and the input bias nulling can bring the cartridge current to well under 1uA (worst case).
In fact, if you connect a 2k cartridge to this preamp, you deserve the "horrible hiss". The noise will indeed be horrible because of the input current noise (wich doesn't cancel as the bias, being from uncorrelated sources). I can estimate the input noise current to about 5pA/rtHz which with your 2k cartridge will give and extra 10nV/rtHz, to add to the 2k impedance noise of 6nV/rtHz plus the 1-2nV/rtHz voltage noise of the preamp. Total is 12nV/rtHz, which is indeed a "horrible hiss" even for a MM 2k cartridge. A decent JFET input opamp like the OPA827 has only 3nV/rtHz typical.
The Paradise should be used only with low impedance MC cartridges. I have to agree, the HF stability of the Paradise is indeed questionable. A small cap (100p or so) connected across the input may cure some stubborn oscillations, a MC cartridge won't care about the cap.