Aleph ONO working - belated thanks to pquadrat

[GaryB]

Originally posted by ungie
I'd love to learn some details of the SuperIt mods you have performed. I have a mint unit here that blew the main power supply cap some years ago.

Well discussion on modifying a NYAL Super It doesn't really belong in the Pass forum. Here are some brief comments. Further discussion should move elsewhere.

Failure of the power supply caps is a classic failure mode in the Super It. It used Photoflash capacitors which tend to die early. Replace them with the biggest Panasonic TSHA capacitor that fits the board pin spacing. You should also replace the diodes with Hexfreds or other good quality low noise diode. I'd upgrade the regulator op amp to an OPA2604 and put a 0.1uf polypro cap across the filaments right at the tube sockets. This is a good start to resurrecting the super it. You can also upgrade coupling caps and riaa caps to good effect. Then if you want to go truly overboard you can replace resistors in the riaa and the cartridge loading circuit.

---Gary

[Stefanoo]

Gary,

sorry for the delay on my reply.
I will quote something and answers are inserted below

Originally posted by GaryB


I've never listened to naked Vishay resistors but I suspect they will be fine. Regarding the RIAA capacitors, I don't think that Auricaps come in the right values. I was looking at the Partsconnexion website and see that Relcap has Teflon 0.01uf and 0.033uf 200v caps that would be excellent. They are expensive but I would recommend that expense. If budget is limited then I'd spend money on the teflon caps before the expensive resistors.
For the 2x220uf caps, I agree with Analog-sa that the Black Gate 220uf 6.3v NX capacitor is a much better choice.


The voltage of the LED is much less than 5v, so you will need to adjust the values of the resistors in the emitters of the current mirror transistors to get the right currents. I don't think this is worth the trouble since LEDs are already low noise.
In the bootstrap cascode, you can not use the AD586. You need to use a two terminal device such as a zener diode and the LM336 that I used is a very good low noise zener. Don't change this.


Yes, that is a typing error. It should have been R67.

Regards,
---Gary

YEs, even Aura Teflon would be very nice, even a better choice than Relcap.
the only problem is to find a 2700pF available.
I can easily find the 2700pF Auricap Cap.
I would be interested on Aura Teflon, if anybody can help me out to find the 2700pF i would love to use these caps.
Of course, PartsConn gives a very good price on these caps.
Expense is also a matter of concern.


For the led: I know you need to re-adjust the current source mirror, of course, to properly set the current on the diff input and on the output stage.
The point is that i think that a monolithic IC would be better than a LED and the expense is not that much (of course 2 good IC can be like 100 LEDS...so there you are )

I don't catch the fact the 2 pins voltage reference subject.
Every voltage reference would perfectly fit there.
I you just need an input voltage to feed it with and you can use the output to properly bias the cascode.
Am I saying something wrong?
I disagree with the quality of the LM336.
I think that a better choices might be achieved, but i do agree yet, it is a decent voltage reference.

Originally posted by GaryB


Martin,
Thanks for the comments.
I respectfully disagree. If the current is changing between the two halves of the differential pair then the voltage of the common sources has to change. It changes less than the signal level since currents in jfets have a square law dependence on the voltage change, but it does change. We can argue whether or not this effect is significant or not but my experience has been positive with the use of the bootstrap cascode configuration so my experience is that this effect is important.




I never experimented with this. The datasheet of the 2SK389 shows that even 5v is plenty to put the transistors into the saturation region, which is the goal. But they can handle higher voltage and it would not hurt. I suspect the effect won't be significant.

---Gary

If you look at the curve you have just posted you will notice that non-linearityes decrease by increasing VDS and a good spot would be something around 8-10V.
I have also read various experiences on cascoding and good results, soundwise speaking are always achieved with a VDS over 7V.

Just my two cents.

Anyways, thank you very much Gary for your useful post.
I think this is a very additive subject for who wants to build a ONO.



Just one stupid question on cartridge load's matching:

I am going to use a Denon DL 103.
It is a 0.33mV output pick up with a suggested load of around 100ohm.
The question is:
when a MC pick up is used do I need to set the capacitance at the input of the MM stage?
If i decide to only use the MC, can i get rid of the input capacitance selection and only use the resistance selection at the MC input's stage?

[Stefanoo]

Gary,

i have just noticed on the picture of you ono posted in this section, that you are using Zen Film caps.
It it right?
Are you using them as coupling cap for the MC stage?

I have been using them on my other implementations and i have been really pleased with the sound.
Did you have a chance to compare them with other film caps and formed an opinion?

I would have loved to use them on the equalization network but they don't come in the right values so.

[analog_sa]

Originally posted by Stefanoo
a suggested load of around 100ohm.
The question is:
when a MC pick up is used do I need to set the capacitance at the input of the MM stage?

100ohm is ok and you may also experiment. I use 470 but it's not very critical. MC carts are not particularly sensitive to loading capacitance anyway but loading the output of the MC step-up is plain silly.

[Stefanoo]

Right, 100ohm or higher (even better) is the suggested load for the DL103.

Got it: capacitance load in not a matter of big concern for a MC pick-up.

This phono linked below (which is the one i have been owning for quite a bit) uses a 1.1nF cap at the input of the MC stage.

http://www.trichordresearch.com/dino.html
(hit technical specs section.)

The ono, instead, has 100pF i guess to prevent RF to get through, but:
Have you ever played with the value of this cap?

Last thing and I actually didn't find this information around on other 3ds (ether i missed it or this thing makes no sense).
Anyways, with regard to the output stage of the MM section:

should the 2 push-pull class A biased output devices 610/9610 (Q18/7) be symmetric i.e. complementary matched to achieve optimal performance?

[GaryB]

Originally posted by Stefanoo
YEs, even Aura Teflon would be very nice, even a better choice than Relcap.
the only problem is to find a 2700pF available.

Stefano,
The Relcap TFE (polystyrene film and foil) caps are very nice and will provide the right values to finish the RIAA network. I still recommend the Teflon caps where you can find them but you can use the Relcaps to finish the network in case you don't find all the teflon caps and you will still have a very good result.

----Gary

[GaryB]

Originally posted by Stefanoo
I don't catch the fact the 2 pins voltage reference subject.
Every voltage reference would perfectly fit there.
I you just need an input voltage to feed it with and you can use the output to properly bias the cascode.
Am I saying something wrong?
I disagree with the quality of the LM336.
I think that a better choices might be achieved, but i do agree yet, it is a decent voltage reference.

A two terminal voltage reference just has an anode and a cathode.

A three terminal device has Vin, Vout, and Ground. This type of device doesn't work for a bootstrap cascode since one wants the Vout / Ground terminals to float up and down with the signal. The Vin terminal makes it hard to move the other 2 terminals with the signal. As I said earlier, you must use a low noise zener for the bootstrap cascode. A low noise voltage reference will note work.

---Gary

[GaryB]

Originally posted by Stefanoo
I am going to use a Denon DL 103.
It is a 0.33mV output pick up with a suggested load of around 100ohm.
The question is:
when a MC pick up is used do I need to set the capacitance at the input of the MM stage?
If i decide to only use the MC, can i get rid of the input capacitance selection and only use the resistance selection at the MC input's stage?

These days I strongly prefer step up transformers over an active MC stage. I recommend you just get rid of the MC stage in the ONO and only use the MM stage. Then you can use a transformer to step up the signal from the DL103. It will sound very good and avoid all the issues with the quality of the ONO MC stage.

---Gary

[analog_sa]

Originally posted by GaryB


I recommend you just get rid of the MC stage in the ONO and only use the MM stage.

This is how i use the Ono at present. From all the cheap transformers i've tried i only liked the Tribute and the amorphous Lundahl. Transformers are another contentious mine field - lots of opportunity to make the wrong choice. I find active step-ups safer and cheaper and in some aural aspects - better.

[Stefanoo]

Hi Gary,


lets start from the caps.
TFE Relcap are not as good as the Auricaps and that's for sure.
The Teflon Caps might be better.
Unfortunately they don't come in all the needed values (2700pF) and i would rather use all auricaps rather than using 2 better caps and a cheap one.
The sound will be almost determined by the worse cap you have on your network.

If i would find all Teflon caps that this would be my choice.


For the bootstrap, i understand what you are saying about the variable ground.
The voltage reference would have to refer to the voltage variable source of the differential.
But what if the ground would be taken on the source (for both Vin and Vout?) ...i don't know if this would work.

Anyways the LM336 is a nice low noise zener it just comes in 2.5V or 5V output.
how about LM4040 by texas.
It's a 40ppm low noise zener.

About the MC step-up:

by getting rid of the inverting buffer stage and thus of the 2 coupling caps and recuperating the absolute phase at the cartridge's clip, i think that this is a nice and simple MC stage which quality depends upon the quality of the parts chosen so:
coupling cap, resistors, matching of the n-jfet, resistors, power supply topology.

If the above parameters are kept enough high then a good result will be achieved.
Just my two cents.

In the next weeks i will be making my interpretation of the ONO and post all the impressions, details and pics to share my experience with you guys.