Hi jeff mai,
all is clear except of L2 impedance (at full volume). You driving your L2 from 160 Ohm impedance source (R1+R2->L1_1:1=R1+R2=160). However After my i/v transformers you will get a high-impedanced output (4-10k), so better to understand what is the receiver (L2) load will be. Usually it is a 10k autoformer, then the output after L2 should be about 3-3.5Vrms at full scale with 10k output impedance.
If the next stage is a low-impedanced amp/preamp, then anyway I offer a versatile PCBs with jfet buffers installation option to lower the output impedance down to 50 ohms or less.
Thus you just need to chose between larger and smaller trafos based the cost difference, all other circumstations you have described was passed by me within previous experiments, including direct working with TVCs after 9018S+my trafos.
Thanks.
Ivan.
all is clear except of L2 impedance (at full volume). You driving your L2 from 160 Ohm impedance source (R1+R2->L1_1:1=R1+R2=160). However After my i/v transformers you will get a high-impedanced output (4-10k), so better to understand what is the receiver (L2) load will be. Usually it is a 10k autoformer, then the output after L2 should be about 3-3.5Vrms at full scale with 10k output impedance.
If the next stage is a low-impedanced amp/preamp, then anyway I offer a versatile PCBs with jfet buffers installation option to lower the output impedance down to 50 ohms or less.
Thus you just need to chose between larger and smaller trafos based the cost difference, all other circumstations you have described was passed by me within previous experiments, including direct working with TVCs after 9018S+my trafos.
Thanks.
Ivan.
Thanks for the reply, Ivan!
I don't know a lot about L2 - I've had it for over 20 years and didn't know enough to ask many questions at the time of purchase. Checking back through old emails, the only info I have is that it's upwards of 20H inductance across the full winding, which would be well over 100kohms at 1Khz.
I don't know a lot about L2 - I've had it for over 20 years and didn't know enough to ask many questions at the time of purchase. Checking back through old emails, the only info I have is that it's upwards of 20H inductance across the full winding, which would be well over 100kohms at 1Khz.
I've spent some time remembering why I chose 160 ohms and re-familiarizing myself with how this stuff works. It was 8 years ago when I last did any DIY! Hopefully my calculations are correct, haha.
If I use all 8 channels of ES9018 in parallel for each channel (it's a full dual mono DAC), I could use 1:4 step up and 400ohm I/V resistor. It gives me 30% higher voltage output than I have now (I don't need more) and 2.5x the output impedance.
Are any of your offerings suited to use at a lesser I/V loading such as this? Is 32ma too much current on the primary?
Also, in your experience what is the difference in connecting the I/V across the primary instead of the secondary? I think I connected mine across the primary because I thought it needed a ground reference and the Sowter 3575 doesn't have a center tapped primary.
Thanks!
Yikes! The resistors soldered in place are R1 / R2 are 330 ohms. How I read the colour bands as 80 ohms, I don't know but this explains why I couldn't get predicted and actual measurements to agree.
The ES9018 DAC models as 3.05 Vpk in series with 208.5 ohms for one half of the differential output in stereo mode (4 channels in parallel.) This works out to ( 330 ohms / ( 208.5 ohms + 330 ohms) ) * 3.05 Vpk = 1.3 Vrms, which corresponds well with what I measure.
I have the option to use mono mode which parallels all 8 outputs. Doing this halves the DAC series resistance and load resistance for the same output level which would reduce the system output impedance.
My goal is to keep the output level near where it is now (1.3Vrms) while keeping the output impedance at the secondary less than 3kohms.
I would probably be most interested in the larger of your transformers. What are the transformer parameters you recommend for this application? Something similar in specs to the Sowter 1465 would seem to suit my goals and give me a different few options to try, but I would like to hear your views on what suits my needs.
Thanks
Hi!
Inductance measurement depends on the measuring signal level at least. 20H at 1kHz is one story, but it is different in most time as manufacturers usually measuring inductance at 100 Hz. At least the regular (permalloy) cores based transformers losts x10 times their inductance from 100Hz to 1kHz (and continue to drop at higher freqs). My transformers has a +/- linear inductance from LF to HF. However, I am not ready to insist on the superiority of any approach in terms of sound quality, just a remark about inductance measurements.
Let me consider that the current output mode (on to the low load impedance) is much better for this DAC than the voltage output mode (high impedances load). The lower impedance will see the DAC, the better SQ you can achieve from DAC chip itself. It is not a theory, it was proved by me myself on 9018S chip based DACs and by my customers. The lower the impedance seen by DAC, the better the result.
The ES9018S DAC chip in mono mode (all the 8 channels are in parallel) gives you 8*3.9mApp=8*1.4mArms=11mArms of the signal.
If you want to get 1.3Vrms below the 3kOhms, then I suggesting to use:
1. Large transformers with TR (Turns Ratio) of 1:16 (15.75 in real).
2. 16R resistor across the primary, 4kOhm resistor across the secondary. 4kOhm reflects to the primary as ~16R.
Calculations:
16R||16R*0.011mArms*16TR=>
8R*0.011mArms*16TR=1.4Vrms
Output impedance is: 16R*(TR*TR)||4k=16R*248||4k=4k||4k=2k
So you will get the output of ~1.4Vrms an 2k output impedance. DAC will see ~8R of the load.
Hello Ivan,
today I made some tests with your Trafos and the output of my CS4398.
To describe the situation I have done a fast sketch in EAGLE. I did not find the right symbol for the trafo, but it does not matter if the Pins are right now.
My idea was to have a common differential filter between PIN 24 and 23 of the DAC as I have in my CD player . After this filter I connect the trafo.
The output of the trafo is connected to the INP+ and GND of the Buffer stage a La Borberly. Since I have done a preamp with that PCB and I had still a pair I wanted to make a trial for DAC in developing phase. I use them as a buffer stage.
Well I have measured with the oscilloscope the output voltage Vrms and compared to the output of the CD player.
Configuration 1
Trafo connected to TP3 and TP4 and on the secondary a 10k resistor. Result: Vrms is some decibel lower than standard , the sound is thinner and bass are not there.
Configuration 2
Trafo connected to TP3 and TP4 and on the secondary a 3.3k resistor. Result: Vrms is at least the half of the nominal , the sound is thinner and bass are not there.
Configuration 3
Trafo connected to TP3 and TP4 and on the secondary NO resistor. Result: Vrms is at maximum and at least the sound is more full.
There is always a lack of low frequencies compared to the cd Player ( I have almost changed everything there. There is a discrete op amp now, but the 2nd order filter is always in use).
Configuration 4
Trafo connected direct to TP1 and TP2 . No differential filter before the primary.
Finally the sound is there. Same volume in Vrms like in the CD player. Bass are there. With the filter the sound is thinner... It is strage because this filter is present in the CD player before the Op amp stage.
Has someone of you some ideas what I can still try?
At the end it seems I do not need any filter. Of course there is some filtering ion the input of the buffer stage.
Thanks to all.
today I made some tests with your Trafos and the output of my CS4398.
To describe the situation I have done a fast sketch in EAGLE. I did not find the right symbol for the trafo, but it does not matter if the Pins are right now.
My idea was to have a common differential filter between PIN 24 and 23 of the DAC as I have in my CD player . After this filter I connect the trafo.
The output of the trafo is connected to the INP+ and GND of the Buffer stage a La Borberly. Since I have done a preamp with that PCB and I had still a pair I wanted to make a trial for DAC in developing phase. I use them as a buffer stage.
Well I have measured with the oscilloscope the output voltage Vrms and compared to the output of the CD player.
Configuration 1
Trafo connected to TP3 and TP4 and on the secondary a 10k resistor. Result: Vrms is some decibel lower than standard , the sound is thinner and bass are not there.
Configuration 2
Trafo connected to TP3 and TP4 and on the secondary a 3.3k resistor. Result: Vrms is at least the half of the nominal , the sound is thinner and bass are not there.
Configuration 3
Trafo connected to TP3 and TP4 and on the secondary NO resistor. Result: Vrms is at maximum and at least the sound is more full.
There is always a lack of low frequencies compared to the cd Player ( I have almost changed everything there. There is a discrete op amp now, but the 2nd order filter is always in use).
Configuration 4
Trafo connected direct to TP1 and TP2 . No differential filter before the primary.
Finally the sound is there. Same volume in Vrms like in the CD player. Bass are there. With the filter the sound is thinner... It is strage because this filter is present in the CD player before the Op amp stage.
Has someone of you some ideas what I can still try?
At the end it seems I do not need any filter. Of course there is some filtering ion the input of the buffer stage.
Thanks to all.
Hello Michele,
Thank you for your experiments and for detailed description of options you tried.
Well... Transformer itself has LPF function as its bandwidth is limited by construction and internal losses at UHF compared to OPamp.
Don't you tried to remove R4, R6 and 10pF as well, in addition to excluding of R1, R2, C1?
Thank you for your experiments and for detailed description of options you tried.
Well... Transformer itself has LPF function as its bandwidth is limited by construction and internal losses at UHF compared to OPamp.
Don't you tried to remove R4, R6 and 10pF as well, in addition to excluding of R1, R2, C1?
Last edited:
Trafos are just trafos, they simply converts impedances. Sure you can use it as a preamp, as ISO-Max from Jensen.
But trafos are different in terms of their purposes. For example the mentioned by you Iso-Max solutiuon is a wide line with different transformers inside box in terms of their impedances, turns ratio and compatibility with different sources and receivers.
Recently, I am trying to make my PCBs for my transformers universal. There are even receptacles on PCB for simple pluggining to allow the changing of transformers with different turns ratio and impedances without soldering. But turning back to your question, it is better if you will describe your source, receiver and the needed gain matching between them.
Hello Ivan , I had forgotten to connect R23 to Ground... that is way maybe the sound was "thin"...now the floor shakes with the bass....My neighbors are really happy with it ;-)
It is strange that the output stage was delivering anyway a Gain of 1...
I did not try again all the possibilities.
Just tried with and without differential filter with R1 R2 and C1 at the primary. I did not noticed any deterioration or macro changes in the sound.
Hello Ivan,
I'm considering your tranformers... I would like to ask... I'm now using active I/V stage for my PCM56 DAC and just tried MC transformer borrowed from my friend - and it certainly sounds very very good, there is exclusively deep soundstage... but there is certain think, maybe it can be related to bandwidth of transformer. There is little feeling of "slowing" of sound.
What is the main reason why not to use classic MC transformers for I/V conversion, when it can also have 3 Ohm primary resistance and 1:10 turn ratio? What I'm missing in my understanding? Thank you in advance.
Kamil
I'm considering your tranformers... I would like to ask... I'm now using active I/V stage for my PCM56 DAC and just tried MC transformer borrowed from my friend - and it certainly sounds very very good, there is exclusively deep soundstage... but there is certain think, maybe it can be related to bandwidth of transformer. There is little feeling of "slowing" of sound.
What is the main reason why not to use classic MC transformers for I/V conversion, when it can also have 3 Ohm primary resistance and 1:10 turn ratio? What I'm missing in my understanding? Thank you in advance.
Kamil
Hello Kamil.
Interesting question
Well...
Strictly speaking, you can use MC-transformers after DAC, no doubt. But with a high degree of probability you will be limited at low frequencies. My experience is too small and I cannot claim a final conclusion, but I can afford the following hypothesis. If we compare vinyl with digital, then vinyl itself has a weaker bass response. My auditory observations confirm that bass is deeper on digital. The main limitation is the overall resonance of the tonearm cartridge (7-14Hz). In practice, such cases are not uncommon when the installation of an MC transformer with a band extended from below (at low frequencies) leads to an overload of the transformer, despite the fact that technically such a transformer is better on paper. The resonance of the tonearm head is capable of overloading a weak-core MC transformer, despite the high inductance of the primary. Manufacturers of MC transformers are forced to take this into account. Transformers with an extended low-frequency range and a good core are usually not cheap.
Why do I so often focus my attention on the low-frequency range (and both response and low distortion are important)? Because from my practice, the lower the bass response in the system, the more microdynamics and realism in the sound (in the entire range).
But returning to the comparison of MC transformers with transformers that I make. In addition to probability of an obvious difference in winding technique, wire material and core geometry, the core material is most likely the main factor. I suspect that with similar dimensions, "my" cores cope better with increased (compared to cartridge's) current from DAC. Well, and most importantly, this is a very narrow hysteresis loop, which allows you to transmit a signal with low delays, which just manifests itself in increased sound speed (transient response) in comparison with
Looks like an ad, but it is really just attemption for answering for your question
.Hello Ivan,
thank you very much for reply... this sounds that it can be this way. There is certainly limitation of tonearm resonance, which I didn't thought about. So it drags me towards your bigger transformers
I believe, that you did good job winding them. And I'm very curious how they sound.Pity is, that everything in this Hifi world is subjective, and trafos can't be heard before buy
What one person likes, another didn't... if you don't know person who did review, and don't know his preferences and his taste, then review don't give you much. I'm tuning and improving my DAC for 4 years, and I went through many configurations which I didn't liked, in spite of fact that measured performance was the same.One question... $500USD is including VAT or without VAT? Because there would be probably additional charge after importing them to EU...
Even the "truth" in most cases is the someone's subjective opinion, not talking about soundHello Ivan,
thank you very much for reply... this sounds that it can be this way. There is certainly limitation of tonearm resonance, which I didn't thought about. So it drags me towards your bigger transformers
Pity is, that everything in this Hifi world is subjective, and trafos can't be heard before buy
One question... $500USD is including VAT or without VAT? Because there would be probably additional charge after importing them to EU...
I have sent PM.
Here is the direct recording from my DAC to ADC.
Initail source - 44/16 CDDA image (flac-files).
SW player - HQP: ext3/NS5->705.6kHz/16bit->NAA (LAN) protocol
HW player (end point): I2S from BBB (BeagleBone Black SBC) with Pure OS from Pavel Pogodin (ppy)
digital chain (after BBB): Ian's Fifopi->Reclockpi-<Andrea Mori's DRIXO 45/49MHz clock+squarer->my I2S->PCM converter
DAC: PCM1704->my i/v resistors->first trafos gain stage->jfet buffer->second trafos gain stage->jfet buffer
ADC: Benchmark ADC1<-Andrea Mori's DRIXO 24.576MHz clock->USB->Pixel phone->24/96 flac
Initail source - 44/16 CDDA image (flac-files).
SW player - HQP: ext3/NS5->705.6kHz/16bit->NAA (LAN) protocol
HW player (end point): I2S from BBB (BeagleBone Black SBC) with Pure OS from Pavel Pogodin (ppy)
digital chain (after BBB): Ian's Fifopi->Reclockpi-<Andrea Mori's DRIXO 45/49MHz clock+squarer->my I2S->PCM converter
DAC: PCM1704->my i/v resistors->first trafos gain stage->jfet buffer->second trafos gain stage->jfet buffer
ADC: Benchmark ADC1<-Andrea Mori's DRIXO 24.576MHz clock->USB->Pixel phone->24/96 flac
Last edited:
