All transformers come with built-in bandlimiting - both at the bottom and the top ends. The HF limit is determined by the leakage inductance which is a function of the coupling between primary and secondary windings. The closer they couple, the lower the leakage L and hence higher the bandwidth - so top of the pile for bandwidth would be a bifilar (or multi-filar) wound transformer. Down at the bottom are split bobbin types.
Nope, that's an impossibility - to have a frequency response change without commensurate phase shift.
In this DDNF diagram the "TZ" resistor is 1.5kohm if you have a 1000pf across it it will give you a LP at -3db at 106khz, there is also the HP filter (10uf 100kohm -3db .1hz) after it. Which the LP with the PCM1704 filters the noise down to a few uV which is very low, unless the DDNF is noiser than a tripple stack 844. Then there's no need for the additional stage (the colouration of a output transformer) and your buffer would be a better output driver than it as well as it's lower output impedance.
Cheers George
Attachments
Last edited:
An unbalanced connection between DAC and pre/amp introduces colouration too because of common-mode noise, using a transformer avoids that. I prefer the colouration of a trafo to that of an unbalanced connection. I know I'm far from alone in having that preference.
Transformer based filter
No. It is the 4n7 capacitor across the primary that makes it a second order filter. That coupled with it's natural roll off at higher frequencies. Phase shift... I'd guess less then a Sallen-Key filter as that has a lot of shift depending on frequency just by the nature of it's design.
No. It is the 4n7 capacitor across the primary that makes it a second order filter. That coupled with it's natural roll off at higher frequencies. Phase shift... I'd guess less then a Sallen-Key filter as that has a lot of shift depending on frequency just by the nature of it's design.
Richard - thanks for the great explanation on trannie bandwidth limiting and cutting common mode noise.
I wonder if this tiny and affordable tranny rated for just 2mW can handle the power straight out of a PCM1704? What is the power output of a PCM1704?
I wonder if this tiny and affordable tranny rated for just 2mW can handle the power straight out of a PCM1704? What is the power output of a PCM1704?
Hi,
Do you mean you supress the cap in // with the I/V resistor to add the OT in serie ? Do you apply a RC to the OT ?
I saw also some supressed the filter // cap (often polystiren cap around 1 nF) but iversampling their Library to send the glitch at higher non hearable frequencies ! Is it really usefull or snake oïl, please ?
WHat is please your advice for a 1:1 OT for a DAC outputt ? I do have a littlr Lundhzl somewhere on a Philips CD player I don't use too much but it has a RC filtering as <Lundhal advise to load it (don't know if loading is the average word?)...
Do you mean you supress the cap in // with the I/V resistor to add the OT in serie ? Do you apply a RC to the OT ?
I saw also some supressed the filter // cap (often polystiren cap around 1 nF) but iversampling their Library to send the glitch at higher non hearable frequencies ! Is it really usefull or snake oïl, please ?
WHat is please your advice for a 1:1 OT for a DAC outputt ? I do have a littlr Lundhzl somewhere on a Philips CD player I don't use too much but it has a RC filtering as <Lundhal advise to load it (don't know if loading is the average word?)...
Transformer filter
Well... The I/V has a first order filter. Then a Buffer. Then the transformer so actually it is down stream from the I/V. The transformer based filter is from a DIY Audio Forums thread. On the secondary there is a Zoebel. As recommended by Lundahl.
Well... The I/V has a first order filter. Then a Buffer. Then the transformer so actually it is down stream from the I/V. The transformer based filter is from a DIY Audio Forums thread. On the secondary there is a Zoebel. As recommended by Lundahl.
I don't remember the ref exactly, believe it was this one : http://www.lundahl.se/wp-content/uploads/datasheets/1684.pdf ! The Zobel on it was not made by me, I have not the knowledge for that !
do we talk about that : http://www.lundahl.se/wp-content/uploads/datasheets/feedback.pdf ?
do we talk about that : http://www.lundahl.se/wp-content/uploads/datasheets/feedback.pdf ?
I looked at Eldam's datasheet, no mention of leakage inductance. I could make an estimate though based on what the bandwidth is stated as (100kHz, -1dB) with the load at 10kohm. It won't be very accurate because 100kHz looks too round a figure to me....
I get about 8mH based on LTSpice simulation.
I get about 8mH based on LTSpice simulation.
Last edited:
Lundahl LL1690
That is a hard specification to find. Best way is likely with an oscilloscope and test generator. The 4n7 on the primary had to be changed to 3n3 with the LL1690, so not a hard and fast rule. Makes sense to experiment on the secondary side as well taking cable length into consideration. 8 mH is likely close enough as a start. I will be taking a short break from DIY due to another round of surgery. PM me if someone needs a response from me. Probably won't be on here for a few weeks.
That is a hard specification to find. Best way is likely with an oscilloscope and test generator. The 4n7 on the primary had to be changed to 3n3 with the LL1690, so not a hard and fast rule. Makes sense to experiment on the secondary side as well taking cable length into consideration. 8 mH is likely close enough as a start. I will be taking a short break from DIY due to another round of surgery. PM me if someone needs a response from me.
Probably won't be on here for a few weeks. Thanks Abraxalito for this.
One friend said me, OT gives less bass, but I haven't this feeling based on the listening of my Philips SACD which can be electronically NOSed (it's a sort of DSP DAC, was famous in France 10 years ago for its good price and good sound and had been strongly tweaked by enthusiastshilips DVP 5500 -SACD & DVD/CD player)
The bass are very good and have a lot of readness, sound "articulation" ! Certainly the help of the zobel ?
But with the opa shematic I posted above I use no output stage, because I trimm to zero mV the offset at working temperature : my pre is 47K inputt and has a DC inputt cap already).
For the few experience I have with buffer, most of the time the sound change with the supply, the caps, etc... Never ending story ! I'm not sure an opa 627 is better than an AD797 here ! YMMV of course !
Have you all read the old thread about "Best resistor for I/V" ? Does anyone remember what bulk resistor Peter Daniel used for its highly tweaked TDA1543 DAC ? It was a Caddock but I havent the ref in mind ?
George traumatized me with this story of 2n2 cap for the JFET filtering at the outputt of the Pedja 's shematic because I use 1K R instead 1K5 (wasn't able to find 1k5 Rhopoint wirewound when I sourced the BOM). To wait for I planned to test some Vishays non Bulk I have on hand:
- CPF21K5000FKE14 (metal 2W, 1%)
-ACO1W1501JA100 (wirewound 1 W, 5%)
- RS01AK500FE12 (wirewound 1W, 1%)
I know the ww resistor is not the dreamed type of resistor for I/V but I use the Rohpoint as many due to the advices of T. Loesch...
By the way, about the // cap I use styren with tin foil, but has anyone tried silver MICA already ? It gives subjective very good result for some aop when small values are involved but never tried as filter I/V resistor ?
cheers
One friend said me, OT gives less bass, but I haven't this feeling based on the listening of my Philips SACD which can be electronically NOSed (it's a sort of DSP DAC, was famous in France 10 years ago for its good price and good sound and had been strongly tweaked by enthusiasts
hilips DVP 5500 -SACD & DVD/CD player)The bass are very good and have a lot of readness, sound "articulation" ! Certainly the help of the zobel ?
But with the opa shematic I posted above I use no output stage, because I trimm to zero mV the offset at working temperature : my pre is 47K inputt and has a DC inputt cap already).
For the few experience I have with buffer, most of the time the sound change with the supply, the caps, etc... Never ending story ! I'm not sure an opa 627 is better than an AD797 here ! YMMV of course !
Have you all read the old thread about "Best resistor for I/V" ? Does anyone remember what bulk resistor Peter Daniel used for its highly tweaked TDA1543 DAC ? It was a Caddock but I havent the ref in mind ?
George traumatized me with this story of 2n2 cap for the JFET filtering at the outputt of the Pedja 's shematic because I use 1K R instead 1K5 (wasn't able to find 1k5 Rhopoint wirewound when I sourced the BOM). To wait for I planned to test some Vishays non Bulk I have on hand:
- CPF21K5000FKE14 (metal 2W, 1%)
-ACO1W1501JA100 (wirewound 1 W, 5%)
- RS01AK500FE12 (wirewound 1W, 1%)
I know the ww resistor is not the dreamed type of resistor for I/V but I use the Rohpoint as many due to the advices of T. Loesch...
By the way, about the // cap I use styren with tin foil, but has anyone tried silver MICA already ? It gives subjective very good result for some aop when small values are involved but never tried as filter I/V resistor ?
cheers
Last edited:
That is a hard specification to find. Best way is likely with an oscilloscope and test generator. The 4n7 on the primary had to be changed to 3n3 with the LL1690, so not a hard and fast rule. Makes sense to experiment on the secondary side as well taking cable length into consideration. 8 mH is likely close enough as a start. I will be taking a short break from DIY due to another round of surgery. PM me if someone needs a response from me.
I hope you have a swift recovery from the surgery my friend.
I've tried many caps on I/V resistor, including silver mica (Soshin SE99). The mica sounds good, but I think the C0G/NP0 ceramic is even better. I was impressed by this little ceramic cap (just 10 cents each) as the SE99 is bulky and expensive.