Input transformers
Installed a pair of Ontronics input transformers to a fully modded blue/black amp today. I posted about these a while back, they are designed as DAC output transformers to be driven by a source in the 100-200ohm range and a load of 10k or so (typical DAC into pre or power amp). This is pretty much how they are being used here as power amp input transformers. Maybe not ideal but useable I believe. I installed between the attenuator and the amp.
Had a listen and compared to another 3116 amp which is almost identical in PSU, attenuator and board mods. I found it hard to tell them apart initially. I did notice the transformers had a small insertion loss which meant I gave the amp one extra click on the attenuator to compensate.
My feeling after some listening is that the transformers give a smoother, less "in your face" presentation, somewhat more refined. I think this is something that would become more evident with longer term listening and may be less tiring than capacitor input. Time will tell. Unfortunately its not my amp so I will have to see what my friend thinks of the change.
Installed a pair of Ontronics input transformers to a fully modded blue/black amp today. I posted about these a while back, they are designed as DAC output transformers to be driven by a source in the 100-200ohm range and a load of 10k or so (typical DAC into pre or power amp). This is pretty much how they are being used here as power amp input transformers. Maybe not ideal but useable I believe. I installed between the attenuator and the amp.
Had a listen and compared to another 3116 amp which is almost identical in PSU, attenuator and board mods. I found it hard to tell them apart initially. I did notice the transformers had a small insertion loss which meant I gave the amp one extra click on the attenuator to compensate.
My feeling after some listening is that the transformers give a smoother, less "in your face" presentation, somewhat more refined. I think this is something that would become more evident with longer term listening and may be less tiring than capacitor input. Time will tell. Unfortunately its not my amp so I will have to see what my friend thinks of the change.
Attachments
The mods Matt mentions are all pretty easy and yield great results.
As a side benefit, replacing the bootstrap caps made mine run much cooler.
I really like the Sure tpa3116, but I think I like the modded YJ blue/black best.
Remove C7 and C22 on your Sure3116 or if you think Sure is on to something TI and IRF totally missed, replace them with at least filmcaps (they are only 1nF) like STmicroelectronics suggest for AD modulation, anything less could result in LARGE distortion like Trevor Marshall showed on another Sure ampboard that also had ceramics across + and - output (which btw was indeed AD modulation). After listening to filmcaps there remove the filmcaps and find out if TI and all other manufacturers are right or wrong regarding outputfiltering for their full bridge amplifiers not using AD modulation. The difference is still very audible, but some might like the filmcap there, some also like the ceramic adding maybe couple of % in distortion there, like the ~3% distortion solely caused by these ceramics in output Trevor Marshall measured for us, and many warnings on the web elsewhere..
Replace the bootstrapsnubber-capacitor with COG type, and you'll hear exactly the same difference like on YJblue with or without prefilter snubber, so yes Sure3116 has a snubber, no it doesn't function like the ones in TI EVM or any other board I have put the snubber on using COG for 330pF position. Sure3116 I have here had 200pF in one output, ~350pF classII ceramics in all other outputs, I don't think the 1 capacitor being smaller totally changes things when listening, but it might 🙂
damn!!
i fried my b/b board which i already started to mod (pbtl, input caps, dc caps).
then i fried my red board modded to pbtl (it was already a zombie. i didn´t know how it was still running. but i actually don´t know what finally gave it the finishing stroke).
both boards only give a very similar 'hiss-noise' and nothing else.
but:
i have a sure board and a 'wiener' already ordered.
now i think about getting another b/b board just for the joy of modding it.
can somebody who owns these boards list some key-settings like
gain
input capactitance
dc caps
output filters
cap quality
...
?
i fried my b/b board which i already started to mod (pbtl, input caps, dc caps).
then i fried my red board modded to pbtl (it was already a zombie. i didn´t know how it was still running. but i actually don´t know what finally gave it the finishing stroke).
both boards only give a very similar 'hiss-noise' and nothing else.
but:
i have a sure board and a 'wiener' already ordered.
now i think about getting another b/b board just for the joy of modding it.
can somebody who owns these boards list some key-settings like
gain
input capactitance
dc caps
output filters
cap quality
...
?
yes, i´m sure the source(s) are ok. (was the pc front panel 3.5 mm output and it still works).
had a small flash around psu when the b/b went west.
the red board doesn´t work since i was too violent with the soldering iron.
maybe it took too much heat?
i already ordered a new b/b moments ago. 😀
had a small flash around psu when the b/b went west.
the red board doesn´t work since i was too violent with the soldering iron.
maybe it took too much heat?
i already ordered a new b/b moments ago. 😀
So what did you do at the time the psu flashed? I "measured" all kind of things the wrong way while board was powered and sometimes speakers gave loud crack and sometimes the amp switched off LOL but always switched on again LOL
And you have Oscons on YJblue? Replace them. I also did connect DC wrong once, some Nichicon electrolytics got exited a little, grew in size and got very very hot and started to blow steam befor I unplugged, board was OK after Nichicons were replaced, room did smell toxic for some time...so could be ok except DCcaps.
no oscons.
i was going to reproduce the datasheet recommendations and had 220uf low esr electrolytes on top and 100nf and 1nf on the backside.
i was ready with re-replacing the 220uf´s with chemicon kze 470uf which i wanted to compare.
the incident happened when i was ready with the soldering and was confused with reattaching a small dc-coupling-cap-bank with higher capacities i have away from the board.
maybe the cap bank was still loaded and i somehow unleashed a current impuls.
then i removed the heatsink from chip.
i have fried one tpa3123 definitely with wrong polarity and this victim has a visible melted spot close to its vcc-input.
i saw nothing like that on the 3116.
with the removed heatsink i saw the danzz-design already has the 100nf and 1nf cap (a look at post#941 would have been enough.
but guess what:
removing the heatsink i hammered the alu-brick into one 1nf cap.
this might very well be wrecked, too.
i was going to reproduce the datasheet recommendations and had 220uf low esr electrolytes on top and 100nf and 1nf on the backside.
i was ready with re-replacing the 220uf´s with chemicon kze 470uf which i wanted to compare.
the incident happened when i was ready with the soldering and was confused with reattaching a small dc-coupling-cap-bank with higher capacities i have away from the board.
maybe the cap bank was still loaded and i somehow unleashed a current impuls.
then i removed the heatsink from chip.
i have fried one tpa3123 definitely with wrong polarity and this victim has a visible melted spot close to its vcc-input.
i saw nothing like that on the 3116.
with the removed heatsink i saw the danzz-design already has the 100nf and 1nf cap (a look at post#941 would have been enough.
but guess what:
removing the heatsink i hammered the alu-brick into one 1nf cap.
this might very well be wrecked, too.
Maybe I was lucky? I did it to two pbtl boards that same time, so 2 boards survived LOL But nothing was charged while connecting, just started boiling when I connected psu's to power later. Both psu's got real hot too btw.
1nF decoupling function at that distance might be irrelevant, I don't have the 1nF's anymore on any of the boards I made lately, many boards come without 1nF from China too, so you can remove and test without it 🙂 If still you feel like touching anything electronic today...seems like one of those days that everything goes wrong 🙂
1nF decoupling function at that distance might be irrelevant, I don't have the 1nF's anymore on any of the boards I made lately, many boards come without 1nF from China too, so you can remove and test without it 🙂 If still you feel like touching anything electronic today...seems like one of those days that everything goes wrong 🙂
I mostly seem to find modded blue boards here.
But what i I have a red one? The simpe 2.0 with volume control?
What can I do to improve the way it sounds? I would like it to respond a bit quicker, sound brighter. I don't know how to put it exactly, but it kind of sounds a bit "soft" to me. Sounds ok on some tracks, but would like it a bit more accurate on others. Can it be acchieved some how?
But what i I have a red one? The simpe 2.0 with volume control?
What can I do to improve the way it sounds? I would like it to respond a bit quicker, sound brighter. I don't know how to put it exactly, but it kind of sounds a bit "soft" to me. Sounds ok on some tracks, but would like it a bit more accurate on others. Can it be acchieved some how?
Last edited:
Those TPA chips can handle reverse polarity and not get fried. Your caps are probably toasted and replacing may bring back to life.
I have audiophonics 9023 DAC dedicated for raspberry
Audiophonics I-Sabre DAC ES9023 Raspberry Pi / I2S - Audiophonics
and the hifiberry
curiously, i prefer the hifiberry in my system, but with the modded tube amp.
EDCOR 10/10k input trans TTPC10k/10k
After listening to my sample HD tracks. I would not recommend the Edcor input transformer 10/10k with the SURE TPA 3116
Pros: Voices sound smoother
Cons: Youll lose 40-50%+ in volume output. The data sheet doesnt show db loss.
Notes: 15/15k edcor might be slightly better. Cinemag or Jenson are probably the best option.
After listening to my sample HD tracks. I would not recommend the Edcor input transformer 10/10k with the SURE TPA 3116
Pros: Voices sound smoother
Cons: Youll lose 40-50%+ in volume output. The data sheet doesnt show db loss.
Notes: 15/15k edcor might be slightly better. Cinemag or Jenson are probably the best option.
Attachments
Last edited:
The PCW10k/10k which I think is the individual transformer used in the TTPC10k/10k states an insertion loss of 0.5db, that seems fairly average for such transformers. But what you describe is way more than that. How did you arrive at 40-50%? Based on volumepot position?
Ah well, hope you didn't have to pay too much for it - may be a cheap option but its still $s spent.
The Sure doesn't have a pot. The volume was digitally controlled on my comp. I didn't see the .5db on the data insert sheet, so thx Christ for finding that for me. I expected also the db loss to be around that much. Im not sure why the output loss. Checked and doubled checked the connections. If after more due diligence, I do find something wrong somewhere. Ill repost an update, however, if after due diligence there is no improvement then I wont update. It was a $30 part, so...im not too hurt about it 🙂 R&D is fun u know...Based on RHINDs review of the Cinemag input transformer, that might be the way to go. But it will cost about $180 but Cost vs Audio improvement is all what its about.
Last edited: