Re-visiting the input transformer discussion... I'm thinking about giving those Edcor transformers a try, since no one has, and they are cheap. Around post #4080, xrk971 had suggested these Edcor umm-9pc xformers. They are unbalanced, though the subsequent discussion suggested ways of using them to feed balanced signal to the TPA3116.
I was looking at Edcor's offerings, and might the TTPC Series be more appropriate? Perhaps the only problem is that you give up some flexibility, as the impedance is fixed. Whereas with the umm-9pc, you have a little flexibility in impedance.
What's the "right" impedance to use? I assume matching exactly is probably best. The tpa3116's input impedance is dependent on the gain setting:
What's the best impedance to select if one wants to retain flexibility in gain setting on the tpa3116?
Likewise, on the input side of the transformer itself: presumably the "best" is that which most closely matches the output impedance of the source. But, assuming one might change the source, impedance to select to retain the most flexibility with source components?
I was looking at Edcor's offerings, and might the TTPC Series be more appropriate? Perhaps the only problem is that you give up some flexibility, as the impedance is fixed. Whereas with the umm-9pc, you have a little flexibility in impedance.
What's the "right" impedance to use? I assume matching exactly is probably best. The tpa3116's input impedance is dependent on the gain setting:
Code:
MASTER/SLAVE GAIN R1 (to GND) R2 (to GVDD) INPUT IMPEDANCE
Master 20 dB 5.6 kOhm OPEN 60 kOhm
Master 26 dB 20 kOhm 100 kOhm 30 kOhm
Master 32 dB 39 kOhm 100 kOhm 15 kOhm
Master 36 dB 47 kOhm 75 kOhm 9 kOhmWhat's the best impedance to select if one wants to retain flexibility in gain setting on the tpa3116?
Likewise, on the input side of the transformer itself: presumably the "best" is that which most closely matches the output impedance of the source. But, assuming one might change the source, impedance to select to retain the most flexibility with source components?
You can use whatever capacitance you want. It's sort of personal taste. I think when the right caps are used, and the quality of most things are good enough, adding more doesn't create a negative situation. However there can be some reasons to use less. Rhing prefers a little less, for whatever reason his bass doesn't have as much as control with too much.
current limiter Inrush Current Limiters | Mouser
Choose one that is rated for your PSU (maybe slightly higher, they're cheap try two?). It goes before the capacitors, on the positive DC output. (ve+, PVCC, etc)
Last edited:
Great, I was thinking of trying these but postage quoted to NZ was $50 for a pair - thats a killer but I may ask for a cheaper postage method.
I was looking at the https://www.edcorusa.com/wsm10k-10k. These say both input and output can be balanced or unbalanced - others in different series say one or the other (UMM-9PC says unbalanced to unbalanced).
As for impedance, I was not sure but thought the usual rules for matching input and output impedances would apply. My understanding is input impedance should be at least 4 times the preceding stage output impedance. Thus a CD output (say 500ohm) easily matches the 10k of the transformer. The trafo impedance of 10K still significantly below the 30k of the 3116.
OR, as you suggest, do you look to match impedances?
Last edited:
For starters I'd go with the 15K:15K transformer (TTPC15K/15K). Set the amp gain to be 20-32dB depending on your requirements and preference, and add a resistor in parallel with the amp input impedance to reduce the effective impedance seen by the transformer to 15K (resistor only required for 20dB & 26dB gain settings, as 32dB gain setting already has a 15K input impedance).
Stage 2 would be to play with non 1:1 transformers to see if transformer voltage gain plus reduced amp gain offered any advantage (time to look at the TPA3116 datasheet to see how performance varies with gain setting).
So you are saying the transformer impedance should match the input impedance of the amp? If so better to get a transformer at 30k/30k (assuming standard gain on the 3116)?
What about wiring to the 3116 - is it better to have a transformer capable of being wired in balanced mode to match the input of the 3116 or is there no advantage over straight 2 wire connection to the + and - inputs and no earth? Is the impedance of the amp across + and - pins or between each pin and earth (or is that a silly question?)?
Last edited:
Well, I would like to try this mod as well, but I am in France, and shipping is not free 🙁
Is another alternative to educorusa?
I am going to ask Edcor if they would ship via "Priority Mail International®
Small Flat Rate Box" which is $24.75 to New Zealand and much more affordable. Just checked, its the same to France so I will let you know how I get on.
Have dealt with Edcor before buying transformer for tube amps I have built. They are good to deal with and well built transformers.
Last edited:
I am thinking for a tube preamp for this tpa3116,
well... I had a look on yj 3 tube 12AU7,
but don't know what can be the result.
Of course, I know I have to change capacitors,
but what is your opinion or alternative?
thanks all!
well... I had a look on yj 3 tube 12AU7,
but don't know what can be the result.
Of course, I know I have to change capacitors,
but what is your opinion or alternative?
thanks all!
Can't speak to the YJ three tube 12AU7 preamp, but this is what I and a few others are using:
Yuan Jing Audio - 6N3 + 6Z4 Rectifier Vacuum Tube Pre-Amplifier + Transformer [A] - $45.70
I upgraded the tubes to GE5670 and mated it with a modded Danzz board (Oscons, ceramic bootstraps, and 20db gain) and the sound is quite wonderful. But even at 20db the gain is way too high. With my phono preamp, which has adjustable gain setting, it's no problem, but with some cds, the sound will distort at moderate listening volume. It's a shame because it is a really nice hybrid for the money (less the $100 American).
Yuan Jing Audio - 6N3 + 6Z4 Rectifier Vacuum Tube Pre-Amplifier + Transformer [A] - $45.70
I upgraded the tubes to GE5670 and mated it with a modded Danzz board (Oscons, ceramic bootstraps, and 20db gain) and the sound is quite wonderful. But even at 20db the gain is way too high. With my phono preamp, which has adjustable gain setting, it's no problem, but with some cds, the sound will distort at moderate listening volume. It's a shame because it is a really nice hybrid for the money (less the $100 American).
Hi Destroyer OS,
The SMPS I am using is rated at 24.5VDC 5A, I have a cap bank of 8*4700uf after it, is CL50 sufficient ?
Thanks
Where's the chart from? I can barely read it, and don't know what device it's for specifically.
On my screen it's really choppy looking, but if I know the part number or what not I can actually look it up.
Inrush current limiters do have a tiny bit of resistance, so they voltage drop a little bit.
If your PSU is way over the amperage of the 3116, then a dumb load might be useful to keep the inrush current limiter heated up for low resistance. It'll lower the efficiency of course...
The only reason to use one is cheapness of sticking with a cheap PSU. You might get better results having it between ground and the capacitors negative side. That way there's no resistance or fluctuation from the PSU, and just capacitors have a little more ESR. It could be kept down by a parallel resistor of a calculated size (parallel to capacitors). This is a bit more complicated, but the capacitors should discharge effectively. If total capacitance is high, any negative effect may be very small or indistinguishable. You may want to parallel a couple film caps not on the inrush circuit, in order to be very low ESR for the big electrolytics. I could be wrong but I think the inrush circuit would dampen any ringing that happens when mixing big and small capacitors. I'll try to draw up a schematic for this later.
Inrush current limiters do have a tiny bit of resistance, so they voltage drop a little bit.
If your PSU is way over the amperage of the 3116, then a dumb load might be useful to keep the inrush current limiter heated up for low resistance. It'll lower the efficiency of course...
The only reason to use one is cheapness of sticking with a cheap PSU. You might get better results having it between ground and the capacitors negative side. That way there's no resistance or fluctuation from the PSU, and just capacitors have a little more ESR. It could be kept down by a parallel resistor of a calculated size (parallel to capacitors). This is a bit more complicated, but the capacitors should discharge effectively. If total capacitance is high, any negative effect may be very small or indistinguishable. You may want to parallel a couple film caps not on the inrush circuit, in order to be very low ESR for the big electrolytics. I could be wrong but I think the inrush circuit would dampen any ringing that happens when mixing big and small capacitors. I'll try to draw up a schematic for this later.
The input impedance of the amp is between each pin and earth. No need for a transformer with centre tap or dual secondaries - adding an earth connection to the transformer secondary just creates another issue to address (low impedance path from amp inputs to ground, necessitating coupling caps - yet removing coupling caps is one of the advantages of using input transformers).
http://www.diyaudio.com/forums/class-d/237086-tpa3116d2-amp-412.html#post4007334
Thanks, does that mean then, that the impedance across the + and - pins (and assuming earth is not connected) is effectively double the stated impedance - i.e. 60k for 26db gain?
I am using SMSL TA2021 base board and I'm quite happy with it (with heavy mods ofc).
This thread made me wonder about TPA3116 and I ordered YJ black board. It arrived at the weekend.
When I checked the board I saw one of the output emi snubber 4.7ohm resistor (close to the speaker output) is not soldered properly. One leg of the resistor is not soldered at all.
So I solder properly and feed it with 17v DC to check. Now one of those resistors gets really hot in few seconds. I got handfull of 4.7ohm resistors so I replaced that one. No change! Is this means the board is defective?
Resistor is marked as Output emi snubber 4/4 in this picture;
This thread made me wonder about TPA3116 and I ordered YJ black board. It arrived at the weekend.
When I checked the board I saw one of the output emi snubber 4.7ohm resistor (close to the speaker output) is not soldered properly. One leg of the resistor is not soldered at all.
So I solder properly and feed it with 17v DC to check. Now one of those resistors gets really hot in few seconds. I got handfull of 4.7ohm resistors so I replaced that one. No change! Is this means the board is defective?
Resistor is marked as Output emi snubber 4/4 in this picture;
An externally hosted image should be here but it was not working when we last tested it.
