I have never tried them, so I don't know. That said, I wish they would say more about their thick film resistance material. Every thick film resistor I have seen tested showed an order of magnitude higher current noise than thin film, with foil being the lowest. That includes Caddock, so it worries me that they just say thick film and nothing else.
Jac
Dear all, but specially Dario..
Just a thought. These are ATE 3CS in the photo.
Two times 0.68 in parallel would get us the right value.
Less problematic to fit in the available space. Two on top of each other. Two because of the power handling.
Ciao, George
Just a thought. These are ATE 3CS in the photo.
Two times 0.68 in parallel would get us the right value.
Less problematic to fit in the available space. Two on top of each other. Two because of the power handling.
Ciao, George
Attachments
Digital-key stocks a Riedel 20 watt that is thin film composition. May need to yank the Johansson thick film 0.33 out out of my FE and try these.
Sonic results from the manganin resistor may be resolution related. Wonder if input and feedback loop cap selections are no longer optimal. They all work together.
Non-linearities in the output resistor definitely will color the Sonics. I always thought the Cerafine caps were grainy and etched. They seem to work fine at C9. It may be something like a Silmic mates better with the new R3.
Sonic results from the manganin resistor may be resolution related. Wonder if input and feedback loop cap selections are no longer optimal. They all work together.
Non-linearities in the output resistor definitely will color the Sonics. I always thought the Cerafine caps were grainy and etched. They seem to work fine at C9. It may be something like a Silmic mates better with the new R3.
Manganins are running in while I'm working from home.
It seem they're improving but orientation probably has to be reversed, they seems much more balanced this way.
This evening I will repeat again tests and critical listening.
Regarding axial resistors possibly no modification is necessary since they can be mounted vertically and use 5W ones without problems.
It seem they're improving but orientation probably has to be reversed, they seems much more balanced this way.
This evening I will repeat again tests and critical listening.
Regarding axial resistors possibly no modification is necessary since they can be mounted vertically and use 5W ones without problems.
Yes, also me would like to give a hint more support to those manganins, before discarding them fast.
At the end, we had decided to put these, in place of the ATE 5CS, which are very much like those RS-5. Like exactly the same structure (and very good sound).
George had reassumed nicely, should be re-run the integration procedure..
Side (snap) note.. I can't play with C9..
At the end, we had decided to put these, in place of the ATE 5CS, which are very much like those RS-5. Like exactly the same structure (and very good sound).
George had reassumed nicely, should be re-run the integration procedure..
Side (snap) note.. I can't play with C9..
Okkey.. Just one more data point:
When we tried together, we listened parallel to to identically equipped amplifiers, with the difference in the R3, only. So it was not by memory.
But and then.. In a different trial we tried to decide again, but in that occasion the (slightly different) amp with the axials kept on sounding so good that we left them in.
So it reminds me, a bit, the debate about which opamp to use. Counts the assembly as a whole, and tastes enter..
Ciao, George
When we tried together, we listened parallel to to identically equipped amplifiers, with the difference in the R3, only. So it was not by memory.
But and then.. In a different trial we tried to decide again, but in that occasion the (slightly different) amp with the axials kept on sounding so good that we left them in.
So it reminds me, a bit, the debate about which opamp to use. Counts the assembly as a whole, and tastes enter..
Ciao, George
I'm not discarding them... not yet at least
My intention was to support both, eventually.
Isabellenhuette can possibly sound different, I'm convinced that those Mundorf are custom Powertrons.
I'll keep you all updated
Very good, I am doing a similar thing but using sockets, as usual:
Attachments
![IMG_0545[1].jpg](/community/data/attachments/766/766441-05f8f9048e0bc2eda7f869c16615c95d.jpg)
I would put it like this: you have a klipping limit(set by diodes) - at +_28Vpeak, 56Vpp.
That is ~20Vrms.
20Vrms, if the amp is capable to keep up with the current, is 66W on a 6ohm load.
The peak current is 28V/6ohm (ideal..) >>
4,66Apeak. Definitely in the max current range for 3886.
Is your speaker able to support sustained 66W?
For music it would be a crest factor less, ideally..
(who has the patience to follow that.. Let's klip!!!)
Ciao, George
That is ~20Vrms.
20Vrms, if the amp is capable to keep up with the current, is 66W on a 6ohm load.
The peak current is 28V/6ohm (ideal..) >>
4,66Apeak. Definitely in the max current range for 3886.
Is your speaker able to support sustained 66W?
For music it would be a crest factor less, ideally..
(who has the patience to follow that.. Let's klip!!!)
Ciao, George
So... let's take 54W -> 18Vrms -> 25,45Vpeak.
If I well remember the gain of My_Ref is 31db -> Vin-peak = 0,72V
Thanks
Well, things are going in an unexpected direction...On the contrary Dale RS-5 have spot-on timbre and richness, a nice improvement from the Caddocks that suffer from their peculiar, greyish, coloration.
And this with steel end caps...
Maybe boards could be modified for axial R3 instead of TO247...who knows? We'll see.
Dario, I ordered a set of the Vishay Dale RS-5s (after your post) and am running them now. They sound quite nice! I also ran Isabellenhuettes and even though they sound richer and sweeter than the Caddocks they still left a slight veil over the music. These Dales sound amazing although I’ve only run them for 30 minutes or so.
BTW, I ran the Isabellenhuette for about an hours w/o heat sinks and they didn’t even feel warm.
Regards,
Pete
Lets remember that sometimes it's not about a driver, but drivers. I have a set a towers that have 5 drivers, thus the max wattage at any range from this amp wouldn't come close to touching them. But I do get where you were going with that.
George, lets hope those watts are quick, high quality, and satisfying, I'm marked down for 4 kits. I can't wait to hear what these amps are capable of, and I'll have enough kits to be able to experiment and see what changes can be made to help improve this amplifier in the future.
I can't wait to hear what these amps are capable of, and I'll have enough kits to be able to experiment and see what changes can be made to help improve this amplifier in the future.
Last edited:
I just got in on the group buy and am looking forward to building this. I have been building tube gear for the past few years and wanted to try something different.
I used to be a licensed electronics tech in the 70s and can handle the build but have a question.
Are there any changes that optimize the circuit for an 8 ohm speakers? I know how the speaker impedance affects the amp performance but I am not an engineer.
Thanks
I used to be a licensed electronics tech in the 70s and can handle the build but have a question.
Are there any changes that optimize the circuit for an 8 ohm speakers? I know how the speaker impedance affects the amp performance but I am not an engineer.
Thanks
I would say.. as it is, it is already optimized for 8ohm speakers, first of all..
In fact, the high rail voltage, the fact of only one current pump (lm3886), the choice of bypass, the diodes..
All converge for giving the best with higher load impedance.
The current pump (voltage controlled current source) is loaded by the speaker. The speaker converts the current back into voltage, which is then taken for the composite feedback loop.
This way also the loop gain is higher with higher load impedance..
In fact, for lower impedances the double current pump version is more ideal ( original evolution design by Penasa).
The doubled current pump gives a higher 'gm', if we leave R3 invariant. This higher gm then compensates for even the lower load impedances.
Ciao, George
In fact, the high rail voltage, the fact of only one current pump (lm3886), the choice of bypass, the diodes..
All converge for giving the best with higher load impedance.
The current pump (voltage controlled current source) is loaded by the speaker. The speaker converts the current back into voltage, which is then taken for the composite feedback loop.
This way also the loop gain is higher with higher load impedance..
In fact, for lower impedances the double current pump version is more ideal ( original evolution design by Penasa).
The doubled current pump gives a higher 'gm', if we leave R3 invariant. This higher gm then compensates for even the lower load impedances.
Ciao, George
Last edited: