with 2145 is even better than 3+3 2SK170
most likely because actually having 6+6 puny critters with nice effective xconductance
but, as always - is there significant sound difference ......
in any case - it sounds good enough
......it took me time, but I really learned to enjoy in that aspect of Papa's approach - there is always major Spark in any construction, sole reason for justifying package existence, while everything else is made to be good enough; as he (cleverly) said - What should I do next, if I make everything As Best Is Possible?
just tiny part of secret how he is having successful business for all these years
(see, Pa is not just smiles and giggles )
most likely because actually having 6+6 puny critters with nice effective xconductance
but, as always - is there significant sound difference ......
in any case - it sounds good enough
......it took me time, but I really learned to enjoy in that aspect of Papa's approach - there is always major Spark in any construction, sole reason for justifying package existence, while everything else is made to be good enough; as he (cleverly) said - What should I do next, if I make everything As Best Is Possible?
just tiny part of secret how he is having successful business for all these years
(see, Pa is not just smiles and giggles
)
Here's a quick RTA at 1V and 10K load. I'm not sure this graph is terribly useful, as I'm not for sure what impact my measurement gear is having on THD at these levels. Either way, looks and sounds good to me. Unfortunately, I forgot to take measurements on the 1+1 version it replaced, but confident the 3+3 version wins 🙂
Attachments
Hi All,
I've been fiddling around some more with REW, and finally have some plots that look like they have measured things correctly, at least. (Nothing like as impressive as codyt's numbers, however.)
The plots below are for the Iron Pre I posted about above, compared with the BA3 I've posted about elsewhere. All were done using a Focusrite Solo, generating a 1000Hz signal at 0.74V, load on the preamps at about 17k. The various output voltages are in the filenames, controlled with the volume pots on the preamps. I picked 0.2V for both the Iron Pre and the BA3 I'm comparing it with, since it's the volume level I'm listening at, and the other two are at full volume, meaning around 1.4V for the Iron Pre, and about 7V for the BA3.
I thought these looked pretty good, especially the Iron Pre. (I have made no serious attempt to fiddle with the BA3 P3 to adjust the H2/H3 balance at high voltage out. The large-ish harmonics at high voltage there are for another thread, I think.) The big surprise for me was how good the BA 3 is for noise and ripple, given the very simple PSU I used; it's technically CRCLCLC, but the inductors are just low value ones I had lying around, and only included just to see what would happen; in practice it's perhaps more like CRC. The shunt supply in the Iron Pre gives measurably better performance, and would be better still if I got the ripples at (odd) multiples of 60 Hz "ironed out". (Pun intended...)
Here are some questions.
1. What is the most likely cause of the small bumps at 300Hz et cetera? Perhaps ringing in the rectifiers? They're not present in the BA3 (unless hidden in the slightly higher noise level around there) when the testing setup was just the same, so they're not very likely to be artefacts of the measurement.
2. I understand that using a well-designed pcb might reduce noise somewhat compared to doing things p2p, but other than that, what else would be a natural place to look to improve performance? The components I used were all old, and while the capacitors all measured OK (both as marked and compared to each other) they are all 15 years old or so. Any idea how much difference that might make?
3. If I wanted to try to reduce THD where would be best to look? I presume this should be in the buffer and/or transformers, not the PSU? This isn't really a priority; distortion is already reasonable, and mostly second harmonic, and the preamp sounds nice, but I'd like to understand a little better if possible.
This has been a really fun project so far! ZM: Thanks for the original design, and for all the input.
Nigel
I've been fiddling around some more with REW, and finally have some plots that look like they have measured things correctly, at least. (Nothing like as impressive as codyt's numbers, however.)
The plots below are for the Iron Pre I posted about above, compared with the BA3 I've posted about elsewhere. All were done using a Focusrite Solo, generating a 1000Hz signal at 0.74V, load on the preamps at about 17k. The various output voltages are in the filenames, controlled with the volume pots on the preamps. I picked 0.2V for both the Iron Pre and the BA3 I'm comparing it with, since it's the volume level I'm listening at, and the other two are at full volume, meaning around 1.4V for the Iron Pre, and about 7V for the BA3.
I thought these looked pretty good, especially the Iron Pre. (I have made no serious attempt to fiddle with the BA3 P3 to adjust the H2/H3 balance at high voltage out. The large-ish harmonics at high voltage there are for another thread, I think.) The big surprise for me was how good the BA 3 is for noise and ripple, given the very simple PSU I used; it's technically CRCLCLC, but the inductors are just low value ones I had lying around, and only included just to see what would happen; in practice it's perhaps more like CRC. The shunt supply in the Iron Pre gives measurably better performance, and would be better still if I got the ripples at (odd) multiples of 60 Hz "ironed out". (Pun intended...)
Here are some questions.
1. What is the most likely cause of the small bumps at 300Hz et cetera? Perhaps ringing in the rectifiers? They're not present in the BA3 (unless hidden in the slightly higher noise level around there) when the testing setup was just the same, so they're not very likely to be artefacts of the measurement.
2. I understand that using a well-designed pcb might reduce noise somewhat compared to doing things p2p, but other than that, what else would be a natural place to look to improve performance? The components I used were all old, and while the capacitors all measured OK (both as marked and compared to each other) they are all 15 years old or so. Any idea how much difference that might make?
3. If I wanted to try to reduce THD where would be best to look? I presume this should be in the buffer and/or transformers, not the PSU? This isn't really a priority; distortion is already reasonable, and mostly second harmonic, and the preamp sounds nice, but I'd like to understand a little better if possible.
This has been a really fun project so far! ZM: Thanks for the original design, and for all the input.
Nigel
Attachments
These look really good to me. The 300hz could simply be a harmonic from the 60hz ‘fundamental.’ Reducing distortion with BA3 would be a matter of twiddling P3 while watching your RTA measurements. You’ll get to that point where the 2nd is largely nulled, which would probably be the lowest THD.
as far as ripple goes, what’s the proximity of your transformer to boards?
as far as ripple goes, what’s the proximity of your transformer to boards?
regarding ripple in IPre, with proper pcbs and especially reg part laid out properly ( separating all GND returns and routing sense lines separately) no way that Focusrite is able to pick up anything
regarding THD reducing/shaping - BA3 threads are place to seek specific info, what to look for and where to look for, with your measuring setup
I have Focusrite Scarlet 2i2 but, frankly, just unpacked it to test basic functionality ( with cans) and packed it, for later
too busy to change already established routine with EMU1212m I have in my workshop PC
regarding THD reducing/shaping - BA3 threads are place to seek specific info, what to look for and where to look for, with your measuring setup
I have Focusrite Scarlet 2i2 but, frankly, just unpacked it to test basic functionality ( with cans) and packed it, for later
too busy to change already established routine with EMU1212m I have in my workshop PC
Thanks for the input, guys.
I'm not looking to discuss the BA3 here; those plots are only really for the purpose of comparison. I'll post elsewhere about that.
ZM: I'm considering getting pcbs from you, if you have them, and building a better Iron Pre. (Because this one is fun, and a better one would be, well, better!) But I want to experiment with this one first, so I'll PM you when ready.
Meanwhile, your comment on traces and pcbs sounds like you think the PSU ripple is just due to my crude P2P power supply - which wouldn't surprise me Is that right? What about using these old caps? Would snubbers on the rectifiers make any difference, or is the shunt supposed to take care of that?
Also: the harmonics from the mainscare only the odd ones, so I'm guessing the shunt PSU is cancelling out the even ones, like an F5 cancels them out with its symmetric design. Is that right?
Thanks again
Nigel
I'm not looking to discuss the BA3 here; those plots are only really for the purpose of comparison. I'll post elsewhere about that.
ZM: I'm considering getting pcbs from you, if you have them, and building a better Iron Pre. (Because this one is fun, and a better one would be, well, better!) But I want to experiment with this one first, so I'll PM you when ready.
Meanwhile, your comment on traces and pcbs sounds like you think the PSU ripple is just due to my crude P2P power supply - which wouldn't surprise me Is that right? What about using these old caps? Would snubbers on the rectifiers make any difference, or is the shunt supposed to take care of that?
Also: the harmonics from the mainscare only the odd ones, so I'm guessing the shunt PSU is cancelling out the even ones, like an F5 cancels them out with its symmetric design. Is that right?
Thanks again
Nigel
dunno did I post this as teaser or not (I believe I did?)..... but first few remarks:
- considering how long it took Jim and resto f Crew to get IPre to Store (not complaining, I always point on fact that majority of Guys are having real day jobs and that this isn't), and
-considering how this (second) iteration of Ipre is not going to be at all in same price range, it can't be choking factor or direct competitor to Store iteration.....
so, it came to me - why not making it, in first place being pleasure to make something in good ole Lego fashion, without too much brain activity, past pure pleasantry level
anyway, (clumsy few sentences up, I see ... still first morning coffee) here it is, made in full blast Iron Pumpkin fashion - big MoBo, TFT/Logic module and encoder and remote, resistor matrix (2^6) attenuator, SE having 6 or 12db settable, Bal just 6db
when? dunno** ........ Store when? - ask Jim; don't press him ..... calmness is important for his day job
**already have SE pcbs for confirmation phase; 'duino code done (trivia - simplification of IPumpkin code and few itsybitsies addition), need to find time/shave few things from list in front
is it another one, Greedy Boyz needing it - dunno, I don't care ..... it deserves to be made, it was fun, all fields ticked
- considering how long it took Jim and resto f Crew to get IPre to Store (not complaining, I always point on fact that majority of Guys are having real day jobs and that this isn't), and
-considering how this (second) iteration of Ipre is not going to be at all in same price range, it can't be choking factor or direct competitor to Store iteration.....
so, it came to me - why not making it, in first place being pleasure to make something in good ole Lego fashion, without too much brain activity, past pure pleasantry level
anyway, (clumsy few sentences up, I see ... still first morning coffee) here it is, made in full blast Iron Pumpkin fashion - big MoBo, TFT/Logic module and encoder and remote, resistor matrix (2^6) attenuator, SE having 6 or 12db settable, Bal just 6db
when? dunno** ........ Store when? - ask Jim; don't press him ..... calmness is important for his day job
**already have SE pcbs for confirmation phase; 'duino code done (trivia - simplification of IPumpkin code and few itsybitsies addition), need to find time/shave few things from list in front
is it another one, Greedy Boyz needing it - dunno, I don't care ..... it deserves to be made, it was fun, all fields ticked
Attachments
Those boards look cool. I can wait...
Meanwhile, I always loved Lego when a kid, and it seems I stiil do!
Maybe I'll try rebuilding the good gemini with a more careful p2p layout and see what difference that makes. (More care with grounds and senses...)
Some marginally more acceptable case might be good, too.🙄
Meanwhile, I always loved Lego when a kid, and it seems I stiil do!
Maybe I'll try rebuilding the good gemini with a more careful p2p layout and see what difference that makes. (More care with grounds and senses...)
Some marginally more acceptable case might be good, too.🙄
to address few things ( I ether didn't saw or you edited while I was typing previous one):
- forgot to add ( even if I did repeat that oomphteen times) - this thread's IPre is made for DiyA Store - and I don't have any saying or anything else (nor judgmental, nor decision wise, nor commercial) with it, except helping G.Boyz when it is needed
- if reg is working properly, ad I'm sure this one is working properly when made as I laid it out - there is nothing to be cancelled with antiphase effect between rails; so , weed you're seeing is due to your actual build so be happy ..... you made it functional, and even if it's made in crude veroboard form, you can enjoy in it for years
- bad caps or not - I'm not sure that snubbing will do much - effect you're seeing is most likely form fact that your sense lines are not so well isolated from energy lines in reg, so cap pumping currents are also affecting sense lines
so - be happy with what you already have - and - whenever you see IPre in Store, then just think (trice) - do you still hear any deficiency in your build and is it worth building it Kosher, once more
Hi ZM,
Thanks again, your input is much appreciated. As a regular diyaudio store customer I'm also grateful for your support for the store, as I'm sure we all are.
Best
Nigel
Thanks again, your input is much appreciated. As a regular diyaudio store customer I'm also grateful for your support for the store, as I'm sure we all are.
Best
Nigel
got logic module pcbs today, will check them soon
then need to populate MoBo and check it
dunno, maybe a month time, maybe less
but- you already have one (as far I remember) , why wasting money on Kosher iteration?
then need to populate MoBo and check it
dunno, maybe a month time, maybe less
but- you already have one (as far I remember) , why wasting money on Kosher iteration?