Since the opamps died what you've done is remove both opamp stages - the first one (the I/V stage) was replaced with a resistor (passive I/V) and the second one was completely bypassed. As a result of bypassing the second stage you no longer have NOS droop correction so there's now a HF roll-off.
Since the opamps are assisted by discrete output stages (unity gain buffers) you're now using one of those discrete buffers after the passive I/V stage. The second buffer isn't being used at present but I could look into re-purposing that to restore the HF response.
Since the opamps are assisted by discrete output stages (unity gain buffers) you're now using one of those discrete buffers after the passive I/V stage. The second buffer isn't being used at present but I could look into re-purposing that to restore the HF response.
I'm playing with a first prototype of an evolution of Kubelik where the all the opamps are removed completely and the analog stage just consists of discrete transistors. Need to run the sound by wifey to see if there are any snags with this.... Seeing as the OPA2209s have shot up in price of late this could be a way to keep the BOM cost in check.
The tricky bit here is managing the DC level - as the transistors all lose 0.6V we have 2.4V between the input of I/V stage and output of droop correction stage. Which reduces headroom rather but I'll have a play to see what might be achieved. With passive I/V there's no need for the 20V power supply of Kubelik, it might be possible to get down as low as 12V.
Oh forgot to mention, the inductors need to be 11mH (previously 22mH). I looked in Mouser and they don't have that value but they do have 12mH so perhaps I can tweak something...
The tricky bit here is managing the DC level - as the transistors all lose 0.6V we have 2.4V between the input of I/V stage and output of droop correction stage. Which reduces headroom rather but I'll have a play to see what might be achieved. With passive I/V there's no need for the 20V power supply of Kubelik, it might be possible to get down as low as 12V.
Oh forgot to mention, the inductors need to be 11mH (previously 22mH). I looked in Mouser and they don't have that value but they do have 12mH so perhaps I can tweak something...
Last edited:
If you'd like to try this variation based on Kubelik you'll need to source yourself some 12mH inductors - that's a value that goes with an I/V resistor around 1kohm. Standard Kubelik uses 2kohm and 22mH inductors. I'm fairly sure it does sound better than original Kubelik - there seems to be more tunefulness in the bass and a deeper soundstage - but so far I've not done the side by side comparison.
I've now built two and running them balanced is another (small) step up.
Wifey hasn't complained about anything so far but she hasn't given it a concerted listen, just the stereotypical listen from a neighbouring room
Here's an interim schematic, I'm giving this evolved design a new name, Dorati (though I often can't remember his name correctly and call it 'dorito'). Not all the component values are adjusted for the lower supply voltage yet, but it'll give some idea.
I've now built two and running them balanced is another (small) step up.
Wifey hasn't complained about anything so far but she hasn't given it a concerted listen, just the stereotypical listen from a neighbouring room
Here's an interim schematic, I'm giving this evolved design a new name, Dorati (though I often can't remember his name correctly and call it 'dorito'). Not all the component values are adjusted for the lower supply voltage yet, but it'll give some idea.
Last edited:
Just a memo.Wifey hasn't complained about anything so far but she hasn't given it a concerted listen, just the stereotypical listen from a neighbouring room
The role of Abrax's wife in the development of his DACs isn’t that stereotypical.
To his credit , Abrax was cautious enough not to make reference to the evil word “kitchen”.
That may save his pants from catching fire.
George
Here's an example of a 12mH inductor which fits the PCB : https://www.mouser.com/ProductDetail/Bourns/RL181S-123J-RC?qs=3lm6YO6aFp4wT7lFi1NfMg==
Great to see so many developments... as ever
Out of curiosity, did you try decoupling some of the "PS" lytics to "see" if it made any difference to the sound?
I am thinking mainly C1, C2, C3 and C4... with MKS or MKP (< 1uF), if desesparate re cost and size enventually X7R caps
Just thinking out loud looking quickly at the posted schematic
Claude
Out of curiosity, did you try decoupling some of the "PS" lytics to "see" if it made any difference to the sound?
I am thinking mainly C1, C2, C3 and C4... with MKS or MKP (< 1uF), if desesparate re cost and size enventually X7R caps
Just thinking out loud looking quickly at the posted schematic
Claude
The 12mH inductor can still be found, but comparing the two schematics I don’t think I will try to pull off this conversion on the current Kubelik board. Are you planning on making Dorati boards and will they also be offered as a kit maybe? I will continue to build the Kubelik as was planned. Thanks Richard – I will PM you in a week or so regarding TDA8932 amp kits. I am very excited about those kits too.
Hi Richard. I have a Kubelik kit already that I've not touched yet but my confidence for the SMD soldering is now at a place where I will start assembling it soon. However, I do like to go off the beaten track and I'm wondering about using an alternative output arrangement based on two Kubelik's in balanced mode. I have all the key parts to build this output stage;
http://itsonlyaudio.com/ultranalog/cdenhancer/cdenhancer2_3a5.pdf
thoughts?
Cheers
Ray
http://itsonlyaudio.com/ultranalog/cdenhancer/cdenhancer2_3a5.pdf
thoughts?
Cheers
Ray
Hi Ray - you can have Kubelik drive trafos directly without needing valves. But if you want the warm glow then by all means try it.
If I were you I'd build the Kubeliks without the opamps, with passive I/V and feed them with an I2S splitter. But you might not want the hassle of building the splitter, that's several logic chips. I'll eventually get on to making a PCB for my splitter, after we've done Dorati's PCB. When wifey built a balanced DAC-amp with a pair of Kubeliks (no splitter, just invert/non-invert) and later generation monoAMPs we didn't think adding the extra DAC was worth the effort in the end for the improvement in SQ.
@commstech - yes the mechanical footprint's going to stay the same.
If I were you I'd build the Kubeliks without the opamps, with passive I/V and feed them with an I2S splitter. But you might not want the hassle of building the splitter, that's several logic chips. I'll eventually get on to making a PCB for my splitter, after we've done Dorati's PCB. When wifey built a balanced DAC-amp with a pair of Kubeliks (no splitter, just invert/non-invert) and later generation monoAMPs we didn't think adding the extra DAC was worth the effort in the end for the improvement in SQ.
@commstech - yes the mechanical footprint's going to stay the same.
Thanks Richard. The circuit I linked to is called the 'CD Enhancer' after all...😁
Might a pair of Celibidache modules be a better option, I would only need the DAC boards without the plug in filters? That way I could still use the Kubelik with my MonoAmp as I originally planned.
I can't quite follow what you're thinking of doing with Celibidache modules - how are they going to work without filters? If you are interested in balanced Celibidache I have some prototype boards in my in-tray to try that out, but again it relies on the I2S splitter so that'll be a couple of weeks away at best.