Hi,
It should be much less than 10 mOhms if it is a well made switch.
The gold-flashed silver contact Elma miniature "01" series is 5 mOhms as a guide, and the NSF "JMK" series is <3 mOhms.
Even a cheap (less than 1 Euro) Lorlin "CK" series with brass/silver-plated contacts claims <10 mOhms.
All these examples are 'initially', so they probably do detereorate with use as a Grayhill version suggests 20 mOhm initially, with a max. of 50 mOhm, but I think that is a poor example of switch contacts.
Regards,
It should be much less than 10 mOhms if it is a well made switch.
The gold-flashed silver contact Elma miniature "01" series is 5 mOhms as a guide, and the NSF "JMK" series is <3 mOhms.
Even a cheap (less than 1 Euro) Lorlin "CK" series with brass/silver-plated contacts claims <10 mOhms.
All these examples are 'initially', so they probably do detereorate with use as a Grayhill version suggests 20 mOhm initially, with a max. of 50 mOhm, but I think that is a poor example of switch contacts.
Regards,
I agree
when you have clean metal to metal contact, as intended, the figures are a few milliohms. Note that dirt and oil can affect all metals even gold. Silver Oxide is a conductor (good) but silver sulphide (from atmospheric contamination) is an insulator. Gold is preferred for small signals, low power, silver when there is some current being switched.
Ted
when you have clean metal to metal contact, as intended, the figures are a few milliohms. Note that dirt and oil can affect all metals even gold. Silver Oxide is a conductor (good) but silver sulphide (from atmospheric contamination) is an insulator. Gold is preferred for small signals, low power, silver when there is some current being switched.
Ted
Bobken said:
Copper wire
Higher values will be manganin wire.
My idea is to integrate a volume control into the passive I/V conversion and connect the power amp directly to the low impedance passive output stage.
The whole analog section is already completely free of active devices and standard resistors.
I have rotary switches with 26 steps that gives 0dB to -60dB in 2,5dB steps + muting.
0,03 ohm is -60dB. 30 ohm is 0 dB
Hi,
That's an interesting concept, and I hope that it goes well for you.
If you are not already aware of it, it would be worthwhile trying some Cramolin-based contact cleaner/enhancer, in my experience.
The most 'popular' being made by Caig Labs (www.caig.com) in the USA, and you will see from their website that this product is used by manufacturers in critical locations like aero-space, & telecoms etc.
Their marketing names have changed over the years (Cramolin, Pro-Gold etc.), but the basic ingredient has remained the same for the 25+ years I have been using Caig products. Several high-end audio manufacturers also use it to provide lower-resistance and more reliable long-term contacts which will enhance the sounds.
It also has some lubricating properties, which helps to prevent plated contacts from detereorating as quickly, and this might be of some help with your rather critical requirements here.
Regards,
That's an interesting concept, and I hope that it goes well for you.
If you are not already aware of it, it would be worthwhile trying some Cramolin-based contact cleaner/enhancer, in my experience.
The most 'popular' being made by Caig Labs (www.caig.com) in the USA, and you will see from their website that this product is used by manufacturers in critical locations like aero-space, & telecoms etc.
Their marketing names have changed over the years (Cramolin, Pro-Gold etc.), but the basic ingredient has remained the same for the 25+ years I have been using Caig products. Several high-end audio manufacturers also use it to provide lower-resistance and more reliable long-term contacts which will enhance the sounds.
It also has some lubricating properties, which helps to prevent plated contacts from detereorating as quickly, and this might be of some help with your rather critical requirements here.
Regards,
Why do you need extremely low contact resistance? If the resistors are connected in series (their sum making up 30 ohms), the switch contact will not affect them. It will be in series with the load, i.e. with your power amplifier that has high input impedance, I suppose. So the low contact resistance may not be very important.
8 mA x 30 ohms = 240 mV voltage at the input of your power amplifier, if I got it right.
8 mA x 30 ohms = 240 mV voltage at the input of your power amplifier, if I got it right.
oshifis said:
Hopefully the contact resistance of the switches and also of the IC sockets do not introduce distortion
Anyway, on average listening level the I/V will be 1 ohm or so.
It was originally planned to be balanced 2 x 31,25 ohm / 0,25 V with 1:2 stepup transformer giving 250 ohm and 1 V.
But I will perhaps change to 2 x 0,06 ohm for -60 dB / 62,5 ohm for 0 dB / 0,5V with 1:1 unbalancing transformer and that gives 1V / 125 ohm.
At the moment I have 4 mA SE into 125 ohm and 1:2 transformer which gives 1V / 500 ohm.
Works very good.
I use naked Siemens noninductive manganin wire wound resistors, sound a bit sweeter than Vishay bulk metal foil or Shallx (???) molded noninductive wire wound.
Bernhard said:
8mA is very little current, OK for gold contacts. Have you considered the effect on your circuit performance when the contact resistance of two switch positions may be different by 10 or 20%? You could ask the maker of the switch what is the tolerance for contact resistance. In your application you may need this tolerance to be quite small (what is 1dB in ratio? I think about 5%) and it is not known what tolerance the switch maker will state. I have never seen a switch specification that gave a tolerance for contact resistance, the only thing you see is a general specification of the nominal value, maybe 3 milliohms or 10 milliohms with few details given for current, signal type or voltage. Maybe you are asking for too much performance from a switch that perhaps it was not the intention of the switch designer. But maybe I am being too cautious, maybe you will enjoy the experiment, why not?
Bernhard said:
8mA is very little current, OK for gold contacts. Have you considered the effect on your circuit performance when the contact resistance of two switch positions may be different by 10 or 20%? You could ask the maker of the switch what is the tolerance for contact resistance. In your application you may need this tolerance to be quite small (what is 1dB in ratio? I think about 5%) and it is not known what tolerance the switch maker will state. I have never seen a switch specification that gave a tolerance for contact resistance, the only thing you see is a general specification of the nominal value, maybe 3 milliohms or 10 milliohms with few details given for current, signal type or voltage. Maybe you are asking for too much performance from a switch that perhaps it was not the intention of the switch designer. But maybe I am being too cautious, maybe you will enjoy the experiment, why not?
peranders said:
It must be seen together with the rest of the DAC,
the good thing:
No more preamp, no more voltage divider for volume control, less cables and connectors.
No opamps, no transistors, no tubes, no standard metal film or carbon resistors in the DAC analog circuit. Only a few filter caps and wire, wire, wire.
16 DACs per channel for lower distortion and low 125 ohm output impedance.
Variable I/V resistor keeps voltage on DAC Iout as close to virtual ground as possible, giving lowest possible distortion.
Very low resistor values give low resistor noise, the completely low impedance circuit has high noise immunity, IMHO.
Simpler power supply, no more extra supply for opamps etc. needed.
This design is as pure as possible and sound is so too, already:
2 x 4 parallel DAC chip -> wound wire to ground -> wound wire -> cap to ground -> wound wire -> cap to ground -> wound wire -> cap to ground -> wound wire -> cap to ground -> transformer ( another wire ) -> output connector.
Possible problems:
lowest volume step -60dB = 0,031 ohm
+ 0,005 ohm variation in contact resistance = -58,8 dB
second lowest volume step-57,5 dB = 0,043 ohm
- 0,005 ohm variation in contact resistance = -58,3 dB is very close to -58,8 dB, so the last steps could become very uncertain.
If the variations are that huge. In that case I would use two switches with one 0,062 ohm in parallel with each switch.
Also as mentioned before, I can use 62,5 ohm for 0 dB and 0,062 for -60 dB if I want.
As for the sketch, the rotary switch switches the different I/V resistors from Iout to ground, that's all.
Compare that DAC to a turntable setup:
MC preamp or high ratio stepup transformer + RIAA preamp + preamp. Lots of opportunities to degrade sound quality
But it is not clear why do you switch a variable resistance from I/V output to GND. Also not clear the purpose of the capacitors.
I would connect a fixed I/V resistor (26 resistors in series) to the output of the DAC just like a conventional variable resistance volume control. The switch rotary contact could go to the output transformer. Alternatively, you could create an L-pad for each switch position, then the I/V resistance could be constant. In any case, I would avoid connecting the switch contact in series with the resistors. Connecting it in series with the output transformer will cause less problems.
I would connect a fixed I/V resistor (26 resistors in series) to the output of the DAC just like a conventional variable resistance volume control. The switch rotary contact could go to the output transformer. Alternatively, you could create an L-pad for each switch position, then the I/V resistance could be constant. In any case, I would avoid connecting the switch contact in series with the resistors. Connecting it in series with the output transformer will cause less problems.
oshifis said:
It is the shortest path.
oshifis said:
They form the analog filter together with all the wound wires.
DAC - variable I/V resistor - variable series resistor for impedance matching - CLCLCLC filter - transformer
The analog filter need to see a constant impedance of 31,25 ohm on its input.
oshifis said:
I liked to avoid lots of solder joints in series. But it is true, there may be alternatives.
oshifis said:
Not sure if I understand this one.
oshifis said:
I could use 26 dividers and select each divider by switch , so the switch does see full voltage and current swing. I need another switch for the impedance matching resistor.
I was looking at the same thing.
The only difference was I was going to use a 4:1 step-up transformer as a passive output. For more details, see the RAKK DAC passive output stage.
The advantage is that the contact resistance is reduced by the square of the ratio (in my case by 16).
The advantage in my case is I can have an output with volume control from a current source DAC with a transformer, a switch and a resistor, and the lower the volume, the lower the distortion.
Currently my I/V resistor is 1K, but I chose that because I have a preamp with gain.
Doug
The only difference was I was going to use a 4:1 step-up transformer as a passive output. For more details, see the RAKK DAC passive output stage.
The advantage is that the contact resistance is reduced by the square of the ratio (in my case by 16).
The advantage in my case is I can have an output with volume control from a current source DAC with a transformer, a switch and a resistor, and the lower the volume, the lower the distortion.
Currently my I/V resistor is 1K, but I chose that because I have a preamp with gain.
Doug
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.