Yes, that was my point... but would like the designers to confirm the intention behind, perhaps it is a subsonic filter, perhaps it worked better that way.. who knows...
I have seen the 20k resistor and it is in parallel with the 100k input impedance of the chip amp (and some capacitive load I prefer to neglect but which lower the input impedance further in the higher frequency range). Hence my quick guess of input impedance being say 16k ish, which is rather on the low side but still OK.
Bottom line: unless told by otherwise by the designers of this project who OF COURSE have the last say... I would have rather seen a minimum of 1.6uF input impedance cap, but that's just me!
That would give a LP RC filfter with roughly Fc = 5Hz which is OK as it would lead to -0.26 dB at 20Hz. Say full range for audio purposes.
Would like confirmation about the above, and wether it is perhaps a desired LP filter with a higher Fc than usual by the designers who, again, did a TREMENDOUS job.
TBH, everything on this schematic has been well thought, so I prefer to ask first before even considering any tweak. The only remaining bits would be rather PS decoupling supply related, but let's skip that for the moment.
Claude
I have seen the 20k resistor and it is in parallel with the 100k input impedance of the chip amp (and some capacitive load I prefer to neglect but which lower the input impedance further in the higher frequency range). Hence my quick guess of input impedance being say 16k ish, which is rather on the low side but still OK.
Bottom line: unless told by otherwise by the designers of this project who OF COURSE have the last say... I would have rather seen a minimum of 1.6uF input impedance cap, but that's just me!
That would give a LP RC filfter with roughly Fc = 5Hz which is OK as it would lead to -0.26 dB at 20Hz. Say full range for audio purposes.
Would like confirmation about the above, and wether it is perhaps a desired LP filter with a higher Fc than usual by the designers who, again, did a TREMENDOUS job.
TBH, everything on this schematic has been well thought, so I prefer to ask first before even considering any tweak. The only remaining bits would be rather PS decoupling supply related, but let's skip that for the moment.
Claude
The 1uF input cap was the initial design from Dibya (which may have come from the ST data sheet design). I subsequently suggested 4.7uF, which is why the latest version PCB grew in size to accommodate that. I think that gets it down low enough.
Xmas Amp - Dibya's TDA7293 by Jhofland
Xmas Amp - Dibya's TDA7293 by Jhofland
Many thanks for this confirmation XKR, and, yes, 4.7uF PPP is perfect in every case then and even overkill.
ST could have indeed gone much lower but presumably because they drive their chip directly with its 100k input impedance.
Anyway, that is solved now... moving on to the next topic
Claude
ST could have indeed gone much lower but presumably because they drive their chip directly with its 100k input impedance.
Anyway, that is solved now... moving on to the next topic
Claude
Probably the last question I have, around C12-C14 and C5, C6 and C9, C11. These are if I am not mistaken the important "close" PS decoupling caps.
I am indeed convinced this amp deserves a great PS and the suggested double mono configuration 4 x 15000uF for +V and -V sounds great. Perhaps even a CRC PS...
However, it is always nice to have most capacities closest to the transistor chip pins. Of course that is even more sensitive when ot comes to the small decouling caps, and of course practical consideration make that impossible with big caps.
Here we have 1uF for each of the small 4 caps, closest to the pins. That is for sure OK as small decoupling caps compromise. More important, given the layout of the board, most builders are going to have the capacities banks mounted remotely from the board per se, so quite a few inches away from the chip pins. That means quite some R and L re cables, and so only the 1uF X5R (low quality, why?) and of course then per rail the 2x220uF lytics will act directly localy.
I guess it works well that way and is cheap. I sadly don't own these boards so can't built one myself - I realise my mistake as this would have been a great training project for my next amp! But I would have loved to install some lugs allowing for some tests/experiments with "cap plateforms"... could be nice as there isn't sadly a lot of space.
I would have loved to know how that amp sounds with the recommended (remote) PS, but with instead of the per rail double "1uF X5R + 220uF lytics combo" say an assembly of "A+B+C"
That assembly would have been fiddly due to the board, the idea being to have the shortest distance to the chip pins with the smallest cap.
- "A" could have been a total of say 2000uF LOW ESR 'lytic, deemed enough to cover the immediate chip needs... more so than the meager 2x220uF. Could be 2x1000uF... or 1x 2200uF and making space for "B"
- "B" could have been a non expensive 4.7uF PPP cap, idealy located above "C"... or if not possible see "A"
- "C" could have been a non expensive 0.1uF PPP (or worst case if space is an issue a 0.082uF C0G cap), located where the current 1uF X5R are
I hope I would have been lucky enough to avoid PS resonnances with these values (TBC with a scope) while making sure the amp chip pins benefit from a great local PS (low ESR across frequencies, capacity reserve...).
Worth considering? Or perhaps it has already been considered by the designers and the suggested way turned out to be the best one?
Have fun either way
Claude
I am indeed convinced this amp deserves a great PS and the suggested double mono configuration 4 x 15000uF for +V and -V sounds great. Perhaps even a CRC PS...
However, it is always nice to have most capacities closest to the transistor chip pins. Of course that is even more sensitive when ot comes to the small decouling caps, and of course practical consideration make that impossible with big caps.
Here we have 1uF for each of the small 4 caps, closest to the pins. That is for sure OK as small decoupling caps compromise. More important, given the layout of the board, most builders are going to have the capacities banks mounted remotely from the board per se, so quite a few inches away from the chip pins. That means quite some R and L re cables, and so only the 1uF X5R (low quality, why?) and of course then per rail the 2x220uF lytics will act directly localy.
I guess it works well that way and is cheap. I sadly don't own these boards so can't built one myself - I realise my mistake as this would have been a great training project for my next amp! But I would have loved to install some lugs allowing for some tests/experiments with "cap plateforms"... could be nice as there isn't sadly a lot of space.
I would have loved to know how that amp sounds with the recommended (remote) PS, but with instead of the per rail double "1uF X5R + 220uF lytics combo" say an assembly of "A+B+C"
That assembly would have been fiddly due to the board, the idea being to have the shortest distance to the chip pins with the smallest cap.
- "A" could have been a total of say 2000uF LOW ESR 'lytic, deemed enough to cover the immediate chip needs... more so than the meager 2x220uF. Could be 2x1000uF... or 1x 2200uF and making space for "B"
- "B" could have been a non expensive 4.7uF PPP cap, idealy located above "C"... or if not possible see "A"
- "C" could have been a non expensive 0.1uF PPP (or worst case if space is an issue a 0.082uF C0G cap), located where the current 1uF X5R are
I hope I would have been lucky enough to avoid PS resonnances with these values (TBC with a scope) while making sure the amp chip pins benefit from a great local PS (low ESR across frequencies, capacity reserve...).
Worth considering? Or perhaps it has already been considered by the designers and the suggested way turned out to be the best one?
Have fun either way
Claude
You are welcome to upgrade bypass caps to X7R or Wima MKS, MKP, etc. we chose many of these parts with good value and function to keep the cost down. As you can see from the measurements, it worked out pretty well - beating the factory data sheet specs. One can solder SMT 1uF 100v X7R 0805’s right at the pins.
That's not always true, More capacitance on main board can increase chance of oscillation of amplifier with High Slew Rate.
What's wrong with X5R?
Many thanks XKR and Dibya for your kind reply.
It measures indeed very well and a lot of efforts have been put into this amp, no doubts - great job.
There is nothing wrong here: where ever, whatever.
I mearly suggested some possible paths for upgrading, which could, or not, result in even more positive. I just wanted to know if these paths had already been explored by you (in wich case I wouldn't bother and trust you), or if they weren't yet (in which case it could possibly be worth looking at it). Whatever, you are absolutely right it would cost a few more $ - and on top one has to find out whereas worth or not
X5R caps are per se not as qualitative as PPPs or C0Gs caps, but can indeed be fine (and often ok with Lytics) and swapping caps tech don't promise anything, one has to try to find out.
Regarding the PS bypass mix, as I mentioned indeed one can get oscillations if not careful hence me mentioning an oscilloscope. Having said that, if you chose values with large steps, don't go too small and use the suggested kind of cap tech you may have a good and sain starting point if lucky. One would have to try and measure and listen.
Thanks again for all this: I will follow with a lot of enthousiasm this thread and see how the original unit sounds as baseline, and for sure some will then try some minor variations that could be interesting.
Have fun
Claude
It measures indeed very well and a lot of efforts have been put into this amp, no doubts - great job.
There is nothing wrong here: where ever, whatever.
I mearly suggested some possible paths for upgrading, which could, or not, result in even more positive. I just wanted to know if these paths had already been explored by you (in wich case I wouldn't bother and trust you), or if they weren't yet (in which case it could possibly be worth looking at it). Whatever, you are absolutely right it would cost a few more $ - and on top one has to find out whereas worth or not
X5R caps are per se not as qualitative as PPPs or C0Gs caps, but can indeed be fine (and often ok with Lytics) and swapping caps tech don't promise anything, one has to try to find out.
Regarding the PS bypass mix, as I mentioned indeed one can get oscillations if not careful hence me mentioning an oscilloscope. Having said that, if you chose values with large steps, don't go too small and use the suggested kind of cap tech you may have a good and sain starting point if lucky. One would have to try and measure and listen.
Thanks again for all this: I will follow with a lot of enthousiasm this thread and see how the original unit sounds as baseline, and for sure some will then try some minor variations that could be interesting.
Have fun
Claude
Many thanks Dibya.
I haven't got boards and I am on other amp projects, but my understanding is some fellows went for this kit so I may well join the train a way or the other, and with pleasure
Your power amp development is THE Class A/B kit I do recommend looking at on DIYaudio to get started, that says something...
Tempted, really...
Agains many thanks for all this effort and sharing it - if I get to do test what I have in mind I will of course let you know, for whatever it might be worth
Enjoy music very much
Claude
I haven't got boards and I am on other amp projects, but my understanding is some fellows went for this kit so I may well join the train a way or the other, and with pleasure
Your power amp development is THE Class A/B kit I do recommend looking at on DIYaudio to get started, that says something...
Tempted, really...
Agains many thanks for all this effort and sharing it - if I get to do test what I have in mind I will of course let you know, for whatever it might be worth
Enjoy music very much
Claude
Xmas Shipping List Status Update
Here is where I am so far, I went through my payment list on PayPal again and cannot seem to find payment by kp93300, or was yours to be combined with an Etsy shop order?
I have not received payment from Darp Malone that I can find either.
I may be mistaken, please help me out if I have this wrong.
Here is where I am so far, I went through my payment list on PayPal again and cannot seem to find payment by kp93300, or was yours to be combined with an Etsy shop order?
I have not received payment from Darp Malone that I can find either.
I may be mistaken, please help me out if I have this wrong.
Thanks, V.
With the TDA3255 amp possibly shipping in a few weeks, I may have to switch to building speakers for a while. I just bought 2 Crestron CNAMPX-16x60 amps. I'm also in the middle of another preamp and two MM/MC phono amps. Just about got the BOM scrubbed on the pre. It seems I spend way too much time trying to find that perfect front panel switch that doesn't look like a cheap toggle!
With the TDA3255 amp possibly shipping in a few weeks, I may have to switch to building speakers for a while.
I just bought 2 Crestron CNAMPX-16x60 amps. I'm also in the middle of another preamp and two MM/MC phono amps. Just about got the BOM scrubbed on the pre. It seems I spend way too much time trying to find that perfect front panel switch that doesn't look like a cheap toggle!
I do, but that's for the power. I'm using anti-vandal switches from TE and C&K. They even have the cute Power symbol behind the cap and are two-tone so I can flash the red LED while it is powering up, and then switch to solid green after power-up delay.
What I want the toggle switches for are for the Tone Defeat, IEC Defeat, Subsonic Filter Defeat, and possibly for selecting between MM and MC, although that last one I'll probably leave as a slider on the back side next to the MC gain select. Don't see the need to keep changing cartridges so often that I need a front panel switch. Basically, things other than power that are binary (on/off/yes/no). Maybe even a Mute Switch for the Speaker Protection board I'm thinking switches like on my Sansui AU-717/ TU-717 or even some of the old Pioneer gear from the 70's.
I could add a microcontroller with a Bluetooth interface and write a cell phone app to change it, but I'm a lot faster at soldering in a switch than I am at writing micro code and Android apps.
Hi Dibya,
Trying to catch up...I found an interesting post from mid Jan to X where you said regarding the datasheet:
"and supply bypass was oversized which made the IC to oscillate in 10Mhz range."
To what suppy bypass were you refering to?
I see one 1000uF caps per PS rail (you used 2x 220uF instead), but then there are far more capacities upstream although admittely farther away so ESR gets "tamed"...
Did you see 10MHz oscillations using the .1uF as bypass with 1000uF?
Was it with a given cap and ESR value?
Would you mind sharing your findings regarding this?
No prob if not of course...
Many thanks again
Claude
PS: my current understanding would be the 1uF bypass cap in your schematic was instead rather chosen to tune the sound, perhaps in the midrange... well noting something higher or equal to .1uF was anyway suitable.
Trying to catch up...I found an interesting post from mid Jan to X where you said regarding the datasheet:
"and supply bypass was oversized which made the IC to oscillate in 10Mhz range."
To what suppy bypass were you refering to?
I see one 1000uF caps per PS rail (you used 2x 220uF instead), but then there are far more capacities upstream although admittely farther away so ESR gets "tamed"...
Did you see 10MHz oscillations using the .1uF as bypass with 1000uF?
Was it with a given cap and ESR value?
Would you mind sharing your findings regarding this?
No prob if not of course...
Many thanks again
Claude
PS: my current understanding would be the 1uF bypass cap in your schematic was instead rather chosen to tune the sound, perhaps in the midrange... well noting something higher or equal to .1uF was anyway suitable.
Last edited:
Once there were chips dedicated for a purpose similar like that widespread in TV sets at least in Europe to select station presets. I think, though, that they're obsolete now.
Best regards!
That’s the neat thing about this amp, you can play around with the choice of bypass caps and see if it makes a difference. Also, lots of choices for input caps. Personally, my two budget favorites are: 10uF 35v Elna Silmic II, and 5uF 50v polycarbonate film axials (what I am using now). The PC’s are NOS I got as a gift from Vunce. I don’t know what brand it is - label only has numbers. But they sound great.
Usually, on Class AB chip amps - I don’t find a need for huge rail caps. 220uF to 470uF is enough. A good thick low impedance cable from the PSU to the board is helpful. Some people like to install a bank of low ESR caps on board. Not sure if that makes such a big difference. Measurements of the damping factor at 50Hz to 100H range would be one way to tell quantitatively. That’s where you need the amp to have a good grip on the speaker.
Usually, on Class AB chip amps - I don’t find a need for huge rail caps. 220uF to 470uF is enough. A good thick low impedance cable from the PSU to the board is helpful. Some people like to install a bank of low ESR caps on board. Not sure if that makes such a big difference. Measurements of the damping factor at 50Hz to 100H range would be one way to tell quantitatively. That’s where you need the amp to have a good grip on the speaker.
Last edited: