Just got two amps
on one of them left chanel plays more quet and no bass
where to dig an anybody point star point...
An externally hosted image should be here but it was not working when we last tested it.
on one of them left chanel plays more quet and no bass
where to dig an anybody point star point...
Stu
It seems you just need a shelving low-pass filter between the dac and the TPA3116. See #5 on this page for the schematic and calculations:
Active Filters
In your case if R1 and R2 were 4.7k ohm and C was .047 uF you would get about 6db drop above 720 Hz which might be about what you need
Thanks Scotty,
So this is just inserted in the signal path? So I would need 2 resistors and one capacitor for each channel?
I am a noob but feel confident I could implement this circuit, however, I get stuck for hours when ordering components - there are sooo many choices!
Could you recommend a manufacturer and component type for the 4.7Kohm resistors and particularly the .047uf capacitor?
I'm presuming this will be much cleaner than a dedicated preamp with tone control....
Stuart
So this is just inserted in the signal path? So I would need 2 resistors and one capacitor for each channel?
I am a noob but feel confident I could implement this circuit, however, I get stuck for hours when ordering components - there are sooo many choices!
Could you recommend a manufacturer and component type for the 4.7Kohm resistors and particularly the .047uf capacitor?
I'm presuming this will be much cleaner than a dedicated preamp with tone control....
Stuart
Attachments
Stu
The best cap configuration will depend on how you implement this. The components are cheap, but shipping to your house vs mine may vary a lot by supplier.
I have bought resistors from ebay here... cheap and measure out ok
100x 1 4W Watt 0 25W 4 7 K Ohm Metal Film R Resistor GRS | eBay
and small lots of caps from here... fast and all seem legit
WIMA MKP2 Metalized Polypropylene Caps 047uF 100V 5 x 5pcs | eBay
Those would be good enough for a system using ceiling speakers if you built the filter on a scrap of perf board.
But you may be able to wire this directly on a cheap DAC using radial poly caps and 1/2 watt resistors (thicker leads). The branch with R2 and C2 just goes from the output + to output ground. R1 has to go inline between the last component on the output circuit and the output jack. Could you cut the trace and bridge the cut with R1? Looks do-able on my $8 usb dac like this one:
New Mini PCM2704 Hi Fi USB DAC Sound Card Board ELNA Capacitor for It | eBay
On a Pi B+ this sounds pretty good. Mine dac is several years old and the L-R outs are reversed! Im thinking I may pull off the output jack and solder on a three wire connector for my TPA3116. Your filter would be a snap to implement at that time.
BTW, if 6 DB is too much attenuation, just parallel R1 with another 4.7k and you will have about 3 DB. If you need more than 6 DB, then instead parallel R2 and you will have roughly 9 DB.
The best cap configuration will depend on how you implement this. The components are cheap, but shipping to your house vs mine may vary a lot by supplier.
I have bought resistors from ebay here... cheap and measure out ok
100x 1 4W Watt 0 25W 4 7 K Ohm Metal Film R Resistor GRS | eBay
and small lots of caps from here... fast and all seem legit
WIMA MKP2 Metalized Polypropylene Caps 047uF 100V 5 x 5pcs | eBay
Those would be good enough for a system using ceiling speakers if you built the filter on a scrap of perf board.
But you may be able to wire this directly on a cheap DAC using radial poly caps and 1/2 watt resistors (thicker leads). The branch with R2 and C2 just goes from the output + to output ground. R1 has to go inline between the last component on the output circuit and the output jack. Could you cut the trace and bridge the cut with R1? Looks do-able on my $8 usb dac like this one:
New Mini PCM2704 Hi Fi USB DAC Sound Card Board ELNA Capacitor for It | eBay
On a Pi B+ this sounds pretty good. Mine dac is several years old and the L-R outs are reversed! Im thinking I may pull off the output jack and solder on a three wire connector for my TPA3116. Your filter would be a snap to implement at that time.
BTW, if 6 DB is too much attenuation, just parallel R1 with another 4.7k and you will have about 3 DB. If you need more than 6 DB, then instead parallel R2 and you will have roughly 9 DB.
I did this mod, as described by xrk971
http://www.diyaudio.com/forums/class-d/237086-tpa3116d2-amp-383.html#post3990270
The only difference was that I used surface mount devices, and slightly different values because it was what was available to me.
This mod is on the right side of my board, between the TPA3116 and the output inductors.
Randy
Last edited:
Balanced drive with THAT 1646
Has anybody tried this as an input stage for driving TPA3116D2 balanced:
http://www.diyaudio.com/forums/analog-line-level/220892-1646-balanced-line-driver-pcb.html
- The module can take SE input , 5K input impedance (needs a low impedance source, so most likely not ideal for tube preamps but should be OK with SS); outputs low 50 Ohms;
- Has a gain of 6 dB, so if needed the TPA's gain can be set to 20 to compensate, presuming a "standard" gain of 26 is wanted;
- Outputs balanced signal, DC-free, so no need for input caps on TPA's board;
Any opinions?
Draki
Has anybody tried this as an input stage for driving TPA3116D2 balanced:
http://www.diyaudio.com/forums/analog-line-level/220892-1646-balanced-line-driver-pcb.html
- The module can take SE input , 5K input impedance (needs a low impedance source, so most likely not ideal for tube preamps but should be OK with SS); outputs low 50 Ohms;
- Has a gain of 6 dB, so if needed the TPA's gain can be set to 20 to compensate, presuming a "standard" gain of 26 is wanted;
- Outputs balanced signal, DC-free, so no need for input caps on TPA's board;
Any opinions?
Draki
I looked up the Edcore transformers.
The 10k:600ohm is a better choice than the 10k to 10k option, I believe. It's a 1:100 ratio but it's probably vastly better than a 1:6 ratio, when you don't want to ever be below 1:10.
An input buffer to the transformer would be ideal.
However that being said, the transformer rolls bass off big time under 100hz, and we don't know how much. At this point you'd want some bass speakers on a separate amplifier, creating a whole new situation.
I'm looking at Cinemag. They have a consumer to pro model that doesn't specify an impedance, so it might be good for the inputs. However you'd still be better off with a buffer before the amp.
I've got no idea when/if Cinemags saturate easily.
The 10k:600ohm is a better choice than the 10k to 10k option, I believe. It's a 1:100 ratio but it's probably vastly better than a 1:6 ratio, when you don't want to ever be below 1:10.
An input buffer to the transformer would be ideal.
However that being said, the transformer rolls bass off big time under 100hz, and we don't know how much. At this point you'd want some bass speakers on a separate amplifier, creating a whole new situation.
I'm looking at Cinemag. They have a consumer to pro model that doesn't specify an impedance, so it might be good for the inputs. However you'd still be better off with a buffer before the amp.
I've got no idea when/if Cinemags saturate easily.
My understanding (best check this)....
Transformers don't have a fixed impedance in a circuit. They reflect the source and load impedance according to the turns ratio of the transformer. A transformer with 1:1 turns ratio (600:600 or 10K:10K) will present a load to the source equal to the amp impedance plus the DCR of the transformer primary and secondary. Similarly, the amplifier will effectively be driven by the output impedance of the source plus the DCR of the primary and secondary. A 10K:600 transformer won't necessarily appear as a 10K load to the source, or 600ohm output impedance to the amp.
Bill Whitlock explains all......
http://www.jensen-transformers.com/an/Audio Transformers Chapter.pdf
Transformers don't have a fixed impedance in a circuit. They reflect the source and load impedance according to the turns ratio of the transformer. A transformer with 1:1 turns ratio (600:600 or 10K:10K) will present a load to the source equal to the amp impedance plus the DCR of the transformer primary and secondary. Similarly, the amplifier will effectively be driven by the output impedance of the source plus the DCR of the primary and secondary. A 10K:600 transformer won't necessarily appear as a 10K load to the source, or 600ohm output impedance to the amp.
Bill Whitlock explains all......
http://www.jensen-transformers.com/an/Audio Transformers Chapter.pdf
Remember that a 10K:600ohm transformer probably has a 4.1:1 turns ratio, so your source will need to provide >4.1V of signal for every 1V of signal that gets to the amplifier.
Im curious to know, if there is a discernible difference between the 10k/10k and 15k/15k input transformer ratings. I heard that the 15k/15k transformer voltage might swing higher causing the amp to compress the signal, with little head room, resulting in the sound to be different. That's the reason I went with 10k/10k.
Last edited:
You're correct for 1:1.
However the impedance of the transformer that isn't 1:1 will shift in your favor so that the potentiometer will see it as being a higher impedance, and therefor drive better. (my initial ratios are not correct, sorry, I was rattling off too fast) But anyway, if you've got 30kohm impedance in, and you use the transformer with 4.1:1 then the amp appears at about 120kohm. This allows the use of a volume device that is hopefully compatible with your source.
The question is, if your source puts out enough juice. I don't know the answer to that question.
But I do know a 10k pot isn't "good" sound on my 3116, and it's not the right ratio.
Wow, I think all of this speakers for why a input buffer is a better option, compatible with all volume devices. I'd go discrete. The B1 Mezmerize is pretty awesome.
The Cinemag's are full frequency spectrum and will block any DC you'd see from typical audio sources such as CD player, DAC, etc. I checked with them on that. They replace the input caps and give you a better control over the amp via running balanced. However a buffer is going to make the difference I'd want. Chances are your sources are too high impedance anyway so you'll want a 100kohm volume control, but it wouldn't work well without the buffer.
However the impedance of the transformer that isn't 1:1 will shift in your favor so that the potentiometer will see it as being a higher impedance, and therefor drive better. (my initial ratios are not correct, sorry, I was rattling off too fast) But anyway, if you've got 30kohm impedance in, and you use the transformer with 4.1:1 then the amp appears at about 120kohm. This allows the use of a volume device that is hopefully compatible with your source.
The question is, if your source puts out enough juice. I don't know the answer to that question.
But I do know a 10k pot isn't "good" sound on my 3116, and it's not the right ratio.
Wow, I think all of this speakers for why a input buffer is a better option, compatible with all volume devices. I'd go discrete. The B1 Mezmerize is pretty awesome.
The Cinemag's are full frequency spectrum and will block any DC you'd see from typical audio sources such as CD player, DAC, etc. I checked with them on that. They replace the input caps and give you a better control over the amp via running balanced. However a buffer is going to make the difference I'd want. Chances are your sources are too high impedance anyway so you'll want a 100kohm volume control, but it wouldn't work well without the buffer.
The impedance at the input of a transformer is proportional to the *square* of the turns (or voltage) ratio. So a 4.1:1 transformer with an 30 kΩ output impedance has an input impedance of 500 kΩ.
FWIW, I have a Lundahl LL1684. It's really the output stage of my ES9018 DAC, performing the IV task.
The 1684 is 1+1:1+1.
On the primary side, the two windings are connected and then tied to ground through a 10K resistor. The other ends of the primary windings go to the + and - out of the DAC.
The cinemags are also 1:1.
IMHO, I don't think you want to stray much from 1:1.
Randy
The 1684 is 1+1:1+1.
On the primary side, the two windings are connected and then tied to ground through a 10K resistor. The other ends of the primary windings go to the + and - out of the DAC.
The cinemags are also 1:1.
IMHO, I don't think you want to stray much from 1:1.
Randy
Oops. Today isn't my day for thinking it appears. I even read that today...
Any which way, stock with a lot isn't nearly as good as the amp can be.
https://www.edcorusa.com/Content/Manuals/TTPC10K-600.pdf
Thats the 10k:600 you mentioned, the 15K:15K is the same for frequency response.
I didn't find those. What catagory are they in? These are what I thought you were using.
It's unfortunate that most buffers require capacitors.