hey Guys,
my cd player is a nad c541. and does not have a tube output. im just waiting for the ldr's to arrive and ill start building right way.
ill keep you guys osted on the progress of the build. and im sure there'll be questions coming.
thanks for your help fellas
Kudos to GeorgeHifi for sharing your knowledge with us.
my cd player is a nad c541. and does not have a tube output. im just waiting for the ldr's to arrive and ill start building right way.
ill keep you guys osted on the progress of the build. and im sure there'll be questions coming.
thanks for your help fellas
Kudos to GeorgeHifi for sharing your knowledge with us.
No success with CCS
Hooked the one I cobbled together last week to test. I need to go back over it, one channel read about 200 - 300 ohms over total range, the other about 20K. Must have some wiring problems.
I can see where this powering technique might pay dividends. Hope to get it operating properly soon.
George
Hooked the one I cobbled together last week to test. I need to go back over it, one channel read about 200 - 300 ohms over total range, the other about 20K. Must have some wiring problems.
I can see where this powering technique might pay dividends. Hope to get it operating properly soon.
George
Looks like the Lightspeed Attenuator has a strong Italian following, Andypairo has been vocal in this forum, wish I could understand what's going on, could be good reading, Google translator makes a mess of it.
I wondered why I was getting abnormal amount of orders from that country.
http://www.audiofaidate.it/forum/topic.asp?TOPIC_ID=2576
Cheers George
I wondered why I was getting abnormal amount of orders from that country.
http://www.audiofaidate.it/forum/topic.asp?TOPIC_ID=2576
Cheers George
Dear All,
I want to remote control the volume on my preamp. After the pot I have an 5687 based stage with high input imp. The output impedance of the CD player (5687 ended DAC) is 47k, similar for the RIAA.
I have tried with relays,
http://www.diyaudio.com/forums/showthread.php?s=&threadid=115006&highlight=
it was not better than the black alps.
Now I try to remote control the rotary switch, but there are difficulties.
LDR would be easier to control, but I'm not sure about impedances.
After reading back a lot, I'm still confused if my output impedance is good enough for this or not.
I'm thinking about to use 4 simple DAC , controlled by a uC. I would make a table with adjusting the values one by one for 1dB steps to about -70, than put it into the code.
Please reply what you think
Thanks,
JG
I want to remote control the volume on my preamp. After the pot I have an 5687 based stage with high input imp. The output impedance of the CD player (5687 ended DAC) is 47k, similar for the RIAA.
I have tried with relays,
http://www.diyaudio.com/forums/showthread.php?s=&threadid=115006&highlight=
it was not better than the black alps.
Now I try to remote control the rotary switch, but there are difficulties.
LDR would be easier to control, but I'm not sure about impedances.
After reading back a lot, I'm still confused if my output impedance is good enough for this or not.
I'm thinking about to use 4 simple DAC , controlled by a uC. I would make a table with adjusting the values one by one for 1dB steps to about -70, than put it into the code.
Please reply what you think
Thanks,
JG
Hi all,
the "Italian Job"
is an attempt to extend the lightspeed concept to remote controllable input selection and volume setting.
Using a microcontroller to measure the NSL32-SRx parts "in circuit" (via jumpers) and to set the appropriate current into the LEDs of the optocuplers would allow to use unmatched devices.
We discussed about the best method to achieve such performance with reasonable resolution but I the project has become too complex IMHO and is stuck at the firmware developement.
A similar concept is however replicable, at the expense of using matched devices and/or manual trimming, with some of the most common IR remote kits.
Cheers
Andrea
the "Italian Job"
is an attempt to extend the lightspeed concept to remote controllable input selection and volume setting.Using a microcontroller to measure the NSL32-SRx parts "in circuit" (via jumpers) and to set the appropriate current into the LEDs of the optocuplers would allow to use unmatched devices.
We discussed about the best method to achieve such performance with reasonable resolution but I the project has become too complex IMHO and is stuck at the firmware developement.
A similar concept is however replicable, at the expense of using matched devices and/or manual trimming, with some of the most common IR remote kits.
Cheers
Andrea
Giordano said:
for a "regular" Lightspeed (as the original made by georgehifi and it's clones) I would say definitely NO. But, read on...
that's more or less what we are trying to do here:
http://www.audiofaidate.it/forum/topic.asp?TOPIC_ID=2576
with such a configuration, input impedances other than that of the original Lightspeed (e.g. the much higher ones you'd need) are possible. Perhaps you may consider joining our efforts...
As Paolo pointed out I forgot to show the most interesting features of the project:
-Contactless switching. Relays are used for channel muting.
-Arbitrary input impedance for each input
-R/C features
-No need for matched devices (self-learning procedure)
-Customizable display
-Coffee maker
Cheers
Andrea
The "hardware" features have been discussed and we reached a good agreement about it, now there is need of people with uC programming experience to make the whole project complete.
Any help is appreciated!
Cheers
Andrea
PS In case of a certain interest we might find the time to make a english-speaking thread about the project.
-Contactless switching. Relays are used for channel muting.
-Arbitrary input impedance for each input
-R/C features
-No need for matched devices (self-learning procedure)
-Customizable display
-Coffee maker
Cheers
Andrea
The "hardware" features have been discussed and we reached a good agreement about it, now there is need of people with uC programming experience to make the whole project complete.
Any help is appreciated!
Cheers
Andrea
PS In case of a certain interest we might find the time to make a english-speaking thread about the project.
UnixMan, I have planned to learn Italian, but unfortunately that is just a plan yet. Yet I can not understand that thread.
I'm happy to join as I can. I always share what I make. (for example I have a reliable stepper control program with sony ir receive if anyone interested, however I have a 1 year old daughter and on some weeks I can spend a lot of time with design, on some weeks nearly zero. I can not promish, however I will probably deep in this if I can purchase a good LDR.
I do program microchip pic16 in assembler. I have very small C experience, and very small 18F experience. (however this is nothing for a 16F even).
Regards,
JG
I'm happy to join as I can. I always share what I make. (for example I have a reliable stepper control program with sony ir receive if anyone interested
, however I have a 1 year old daughter and on some weeks I can spend a lot of time with design, on some weeks nearly zero. I can not promish, however I will probably deep in this if I can purchase a good LDR. I do program microchip pic16 in assembler. I have very small C experience, and very small 18F experience. (however this is nothing for a 16F even).
Regards,
JG
Attachments
Andypairo, that looks promishing.
However I have some recommendations :
for sure I will :
- not use relays. I did an experiment with a few type of relays. Omron G6K is quite good, but switch is better. I do not know why. Some people say to use mercury wetted relays. I can't get it for my attenuator. It is true, that 1 relay is used in an input swith and 5-6 in an attenuator network, so it effects more, but I clearly proved for myself at least, that even the better signal relays are not good enough.
- not use fancy displays. During I did the current vol. control, I found if I send the uC to sleep, the sound quality comes back as it is not there at all. When it runs, it is recognizeable. (I have used separate transformer, everything is separated but in the same box.) An LCD display, or a muxed display will have clock or requires the pic to run. I believe the way to shut the uC up, make the button and IR control, on interrupt, that wakes up the pic, it does process the request end goes to sleep again.
Use a 44 pin pic (cheap still, no much difference) and just wire each segment of 7 segment LED displays.
When you listen to music, the pic is virtually not there, all lines are static, no clock (except if you have much IR noise which keeps the pic waking up).
If I do the code for myself, I do it an easy way. I measure the steps and then build into the code. That is easy. I have IR receive routine for sony, have SPI routine for the DACS, all I have to do is to create the 4 table and keeps stepping up and down on the table by the received commands.
In the other hand, the elegant way would be to store the table in the eeprom. Use a two mode program, make a temporary LCD interface and a routine which is finetuneable and it stores the result in eeprom. This way, it would be possible that who build it, does not have to touch the code.
The second one, I can not promish to do, however not a big deal at all.
Regards,
JG
However I have some recommendations :
for sure I will :
- not use relays. I did an experiment with a few type of relays. Omron G6K is quite good, but switch is better. I do not know why. Some people say to use mercury wetted relays. I can't get it for my attenuator. It is true, that 1 relay is used in an input swith and 5-6 in an attenuator network, so it effects more, but I clearly proved for myself at least, that even the better signal relays are not good enough.
- not use fancy displays. During I did the current vol. control, I found if I send the uC to sleep, the sound quality comes back as it is not there at all. When it runs, it is recognizeable. (I have used separate transformer, everything is separated but in the same box.) An LCD display, or a muxed display will have clock or requires the pic to run. I believe the way to shut the uC up, make the button and IR control, on interrupt, that wakes up the pic, it does process the request end goes to sleep again.
Use a 44 pin pic (cheap still, no much difference) and just wire each segment of 7 segment LED displays.
When you listen to music, the pic is virtually not there, all lines are static, no clock (except if you have much IR noise which keeps the pic waking up).
If I do the code for myself, I do it an easy way. I measure the steps and then build into the code. That is easy. I have IR receive routine for sony, have SPI routine for the DACS, all I have to do is to create the 4 table and keeps stepping up and down on the table by the received commands.
In the other hand, the elegant way would be to store the table in the eeprom. Use a two mode program, make a temporary LCD interface and a routine which is finetuneable and it stores the result in eeprom. This way, it would be possible that who build it, does not have to touch the code.
The second one, I can not promish to do, however not a big deal at all.
Regards,
JG
Giordano said:
I guess he meant that the output impedance of such a stage is the anode load (in this case 47K) in parallel with the dynamic impedance of the triode. That is, you're more likely around 1 or 2 Kohm, not 47K!
Giordano said:
I would guess that...
we should resume the important parts in English.don't worry, we do NOT plan to use relays (or any other contact for that matter) on the active signal path -- the design is completely CONTACTLESS!
As Andy said, we are planning to use a relays only during calibrations and for "extra muting" of UNUSED channels. These will be isolated from the signal path by the HUGE impedance of an "open" LDR.
Indeed. We were in fact thinking about exactly something like that!
well, it could be done... but then you loose much of the countless possibilities that an uP controlled LED/LDR attenuator may offer.
And, most of all, to carefully measure the LDRs and create the "steps" to hard code in the end will take you more time than what you'd save from the coding...
this is -more or less- what we were thinking about.
Actually, our idea was to do define a simple "fit" function to the LDR I/R curve, and use an initial automatic calibration procedure to store in the eeprom the fit parameters for each LED/LDR.
That way we can easily set whatever resistance we need to "program in" ANY attenuation curve that we want, with whatever input/output impedance, etc.
It's a bit more work, but it would be the Ferrari (or the Rolls-Royce, if you prefer
) of the attenuators... the possibilities are simply countless!I prefer a Ferrari (or some other from the region , certainly not a Rolls-Royce.
About output impedance I mean the output impedance of the stage, including the capacitor at the output. To go more close to the triode stage out impedance, I would need large capacitor which is not practical and good choice either. If we calculate the output impedance with a 470nF cap, we have good low frequency performance from around 47k. To go under 47k I would have to enlarge the cap.
I'm happy that your concept is nearly the same, however I do not understand the fit function.
You are right, that it takes time to find out the values to the table, the eeprom programming might be easier.
I'm thinking the setting to be done that way :
- calculate the value of the series resistor in ohms for each steps
- put a resistance meter and finetune the drive values to get those resistance values
- calculate output DC voltage for each steps with a given input DC and fine adjust the parallel LDR's drive values to get that voltage leveles
This is a quite long procedure, but I would rather do it for each step, to get a good channel matching performance.
I'm thining of a MAX525 as DAC, a high drive opamp to drive the LEDs.
Regards,
JG
, certainly not a Rolls-Royce.About output impedance I mean the output impedance of the stage, including the capacitor at the output. To go more close to the triode stage out impedance, I would need large capacitor which is not practical and good choice either. If we calculate the output impedance with a 470nF cap, we have good low frequency performance from around 47k. To go under 47k I would have to enlarge the cap.
I'm happy that your concept is nearly the same, however I do not understand the fit function.
You are right, that it takes time to find out the values to the table, the eeprom programming might be easier.
I'm thinking the setting to be done that way :
- calculate the value of the series resistor in ohms for each steps
- put a resistance meter and finetune the drive values to get those resistance values
- calculate output DC voltage for each steps with a given input DC and fine adjust the parallel LDR's drive values to get that voltage leveles
This is a quite long procedure, but I would rather do it for each step, to get a good channel matching performance.
I'm thining of a MAX525 as DAC, a high drive opamp to drive the LEDs.
Regards,
JG
I think you guys who have come over from the Italian LDR discussion can help all of us advance our projects. So far, this thread had been mostly productive and beneficial. My hopes are that it does not get bogged down in the little details that are not important to the improvement of the current delivery to the LDR, etc.
Hi guys, just on a different note, there is one of my commercial Lightspeed Attenuators second hand for sale here in Australia to Australian buyers only, for a saving of 40% on the new price.
It is in mint condition, but being one of the first MkII's it is in a slightly shorter case than the production ones of today, it is exactly the same otherwise.
It is owned by our Audio Club (ASON) President Richard Wong who bought a new one because it matches his other equipment better as far as looks go. He wants just $300aud for it and you can contact him on Sydney 99727139 if you want to ask questions.
Cheers George
It is in mint condition, but being one of the first MkII's it is in a slightly shorter case than the production ones of today, it is exactly the same otherwise.
It is owned by our Audio Club (ASON) President Richard Wong who bought a new one because it matches his other equipment better as far as looks go. He wants just $300aud for it and you can contact him on Sydney 99727139 if you want to ask questions.
Cheers George