| BrianGT |
I am currently working with my roommate on a remote controlled passive volume control. Here is my current, first working version:
http://www.diyaudio.com/forums/attachment.php?s=&postid=94968
This is basically a proof of concept version, consisting of 2 channels with 4 relays per channel, giving 16 total volume steps. The design is based on 8 relays per channel, but there wasn't room on the protoboard for more then 4 for each channel. Currently, only one side of each relay is being used, but with both sides being used, it will act as a 2 channel balanced volume control with 256 steps including mute. Here is the schematic for the relay volume control:
http://www.diyaudio.com/forums/attachment.php?s=&postid=94970
We are currently working on the details of the project, and starting to make a pcb to accomidate all 16 relays, and the microcontroller, a PIC18F452.
Others on this forum have expressed interest in this project, as a lot of people are looking for a high quality volume control. By simply using relays and resistors, it is quite a clean volume control. Significant lengths will be taken to ensure a clean audio path on the pcb.
How many people on this forum would be interesting in making this a group project?
I could supply the finished pcb, relays, programmed microcontroller and control logic in the form of a kit. I thought about just providing pcbs, but relays and other parts are much cheaper when purchased in large quantity, and I will need to program all of the microcontrollers for this project for people anyway.
As for the remote control aspect, my roommate is working on the code for a universal learning remote control unit for it now, so that it will work with any remote control. I like this, because I like to use the unused buttons on my normal remotes for stuff like this.
As for display, I am using this display for mine:
http://www.diyaudio.com/forums/attachment.php?s=&postid=94665
(image from other project with it)
It costs $100, but for the kit, I am considering making it work without a display, with this display, or with a significantly cheaper lcd display (~$10).
As for using it, I am planning on using mine for two setups, one for my normal leach amp, seperate from it, and the other, integrating it in the same chassis as my bosoz.
Anyone on the forum interested in making this a group project, with my supplying of a kit for this? Any comments on what you would like in a passive volume control?
Most of the credit for this project goes to my roommate, who spent a lot of time making the first prototype for a school project this semester.
--
Brian |
|
|
| dorkus |
| not sure it's a practical approach for SoD (son of dork) since i need lots of 1dB attenuation steps, but the control scheme is cool. where'd you get the nice VFD? |
|
|
| BrianGT |
| quote: | Originally posted by dorkus
not sure it's a practical approach for SoD (son of dork) since i need lots of 1dB attenuation steps, but the control scheme is cool. where'd you get the nice VFD? |
Not enough steps? There are 256 steps, including mute with the 8 relays... add 2 more relays, and you have 1024 steps total. Modify the 8 resistors to get whatever resistance that you want.
I ordered the vfd direct from Noritake. Here is the page for this one:
http://www.noritake-elec.com/7000.htm
It was just under $100 shipped.
The model number is:
GU140X32F-7002
It is very useful to use, and has many advanced features over normal vfds. You can set different windows on it.
--
Brian |
|
|
| Dennis Hui |
Hi Brian,
I am definitely interested in this project. One thing I
really like is that this is balance is also handled, but
probably with greater ease of use than standard dual
mono attenuators.
Dennis |
|
|
| dorkus |
| have you figured out the optimal resistor values for precise 1dB attenuation increments? i sure haven't... i wonder if there's a mathmatical solution better than brute force (iterative). |
|
|
| apassgear |
Yes Brian, good idea to make a kit.
As I told you before I'm also interested in a set as volume control for BOSOZ. The idea of adding a display is very nice too.
I'm in for a two channel balanced vol.
Take your time is OK for Jan/Feb.
Since aparently the signal is going through the relays we maybe don't want to cheep-out on these? |
|
|
| dorkus |
| why switch the signal thru the relays? any opposition to a simple shunt control with fixed series resistor? |
|
|
| jackinnj |
| there has been a seller of Noritake displays off and on -- check Businesss Equipment/Electronic Components on the Bay |
|
|
| BrianGT |
| quote: | Originally posted by dorkus
have you figured out the optimal resistor values for precise 1dB attenuation increments? i sure haven't... i wonder if there's a mathmatical solution better than brute force (iterative). |
I haven't even tried yet :) Currently, it is just using 4 resistors for the proof of concept:
2k
1k
500
250
This gives 15 positions plus mute, which works fine to see how it works. I will test the next version with these values:
8k
4k
2k
1k
500
250
125
60
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by apassgear
Yes Brian, good idea to make a kit.
As I told you before I'm also interested in a set as volume control for BOSOZ. The idea of adding a display is very nice too.
I'm in for a two channel balanced vol.
Take your time is OK for Jan/Feb.
Since aparently the signal is going through the relays we maybe don't want to cheep-out on these? |
I am using NEC MR82 relays:
http://www.nec-tokin.net/now/englis...elay_MR82_e.pdf
I will try other types of relays once to see the difference with it. I have no problems with these relays thus far. Here is the technical info for the relays:
http://www.nec-tokin.net/now/englis..._tehcdata_e.pdf
I love the Noritake 7000 series display. It is great looking, and very capable for advanced features and programming.
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by dorkus
why switch the signal thru the relays? any opposition to a simple shunt control with fixed series resistor? |
I am simply following the same path as the Pass Labs P1.7 preamp used for the volume control:
http://www.passlabs.com/pdf/aleph/apserv17.pdf
I have read quite a few reviews of the revision 2 saying that it worked great as a volume control. I see no problems with it. It works great, and is a whole hell of a lot better then the PGA2310.
--
Brian |
|
|
| csd |
I made a plot of attenuation (in dB) vs. steps for the P1.7 volume control - it isn't even close to logarithmic. Note that a logarithmic response would appear as a straight line in this plot. Near 0dB attenuation, the step size is the largest, at about 4dB.
For 8 relays, the maximum possible attenuation is 48dB. The resistor values in the P1.7 design are already optimized - changing any of the values results in step discontinuities in the plot.
Also interesting to note is that it isn't obvious how to switch the relays - it's not just a simple binary count. It took me a while to figure it out. For the first half of the range, you count up from zero in steps of two. For the second half of the range, you count down from 255 in steps of two.
Craig |
|
|
| BrianGT |
| quote: | Originally posted by jackinnj
there has been a seller of Noritake displays off and on -- check Businesss Equipment/Electronic Components on the Bay |
I have seen this, but they are not the newest 7000 series displays from Noritake, which are absolutely great. They have a lot of advanced features that makes them harder to program, but very capable. They have a great windowing feature that allows you to set windows and talk to each of the seperately, so that if you want to just change one thing on the display, you don't have to fool with the rest. This solves the text positioning issues of other displays. The picture above from the alarm clock has 3 seperate windows being used. I haven't programmed it myself, as my roommate did all of the code for this project, but he is in love with it, and wants to purchase one for himself soon, once I take this one back from him to use for myself.
If anyone finds the Noritake 7000 series displays for cheaper then $100, let me know. I was quite hesitant of purchasing such an expensive display, but you really have to see it in person to see how beautiful it is. It is also very easy to see from across the room with the double sized fonts.
--
Brian |
|
|
| dorkus |
cleaner, better sounding, easier to implement. input impedance varies a lot, but who cares.
i think the VFD is worth every penny. just make sure to put it in a separate chassis, or shield it really well, i hear they are quite noisy. |
|
|
| BrianGT |
| quote: | Originally posted by csd
I made a plot of attenuation (in dB) vs. steps for the P1.7 volume control - it isn't even close to logarithmic. Note that a logarithmic response would appear as a straight line in this plot. Near 0dB attenuation, the step size is the largest, at about 4dB.
For 8 relays, the maximum possible attenuation is 48dB. The resistor values in the P1.7 design are already optimized - changing any of the values results in step discontinuities in the plot.
Also interesting to note is that it isn't obvious how to switch the relays - it's not just a simple binary count. It took me a while to figure it out. For the first half of the range, you count up from zero in steps of two. For the second half of the range, you count down from 255 in steps of two.
Craig |
I haven't yet heard 8 relays yet, but by using 4 relays, it sounds alright, just using a simple binary count. Once I get a board done with 8 relays, I will see how the best way to control the software. If you are going by 2 for the count, this makes no reason to include the lsb relay....
This is a work in progress, but I suspect that the P1.7 setup will work fine for my application.
--
Brian |
|
|
| dms |
Your work is interesting, please keep us informed. I am planning a preamp/crossover, but won't begin until I finish the Aleph 4 under construction. What I really need is a multichannel device since my stereo is part of my video system. Also I would like to control back ground music in the house. So I may be looking at 8 to 10 channels with input, gain, and crossover functions.
Don |
|
|
| dorkus |
| don, check out the Son of Dork preamp threads. it's been idle for a while but it outlines the design criteria for a multichannel preamp. my system is hybrid audio/video as well, so i set out to build a multichannel preamp for SACD playback, as well as integration with a surround sound system. |
|
|
| apassgear |
Brian,
In MHO the relays you propose look great (NEC RM82). But in the other hand I have little experiance with these devices.
So from my point of view go ahead whith these ones. You don't need to test others if you find these don't degade the sound. If you do test others (time consuming) I'm sure you would end trying to hear differences the same as if you were trying to decide on which 2" cable to use for a signal. |
|
|
| BrianGT |
| quote: | Originally posted by apassgear
Brian,
In MHO the relays you propose look great (NEC RM82). But in the other hand I have little experiance with these devices.
So from my point of view go ahead whith these ones. You don't need to test others if you find these don't degade the sound. If you do test others (time consuming) I'm sure you would end trying to hear differences the same as if you were trying to decide on which 2" cable to use for a signal. |
Is the DPDT relay footprint the same for most all relays? If it is, that I can make it so that you can input any voltage for the relays to use to trigger on, and then this would allow people to use whatever relays that they want to use.
Does the voltage of the relay make any difference? I am just using 5v relays for the ease of a single supply, but I can make the layout so that you can use any voltage to trigger the relays if desired.
--
Brian |
|
|
| apassgear |
Brian,
In my case at least I will need to make a separate PSU to drive the PIC PCB, so if the relays have the same voltage that what's needed for the PIC board this is excellent for me.
What would be the voltages requiered? |
|
|
| BrianGT |
| quote: | Originally posted by apassgear
Brian,
In my case at least I will need to make a separate PSU to drive the PIC PCB, so if the relays have the same voltage that what's needed for the PIC board this is excellent for me.
What would be the voltages requiered? |
Currently, I am using 5v relays, as that is the voltage that the PIC, display and other control chips. It is obtained from a 9v 1A wallwart, and a LM-340T-5.0v regulator. The VFD draws about 500mA of current, or so it seems, but everything else doesn't draw much. Since none of the power is in the signal path, I didn't figure that it needed a better power supply. Am I wrong?
Here is the schematic for the pic board that I am using (page 6):
http://www.picbook.com/%7Epbpdf001/bookinfo/CA1.pdf
It is from a class here at georgia tech. More details at:
www.picbook.com You can actually purchase pcbs and a parts kit from digikey: 18F452-KIT-ND if you are interested in making your own pic development board. The whole thing could be obtained for about $120 including the book.
--
Brian |
|
|
| apassgear |
| quote: | Originally posted by BrianGT
Currently, I am using 5v relays, as that is the voltage that the PIC, display and other control chips. It is obtained from a 9v 1A wallwart, and a LM-340T-5.0v regulator. The VFD draws about 500mA of current, or so it seems, but everything else doesn't draw much. Since none of the power is in the signal path, I didn't figure that it needed a better power supply. Am I wrong?
--
Brian |
I guess a simple PSU, as you suggest, would be enough. In my case I will do a "formal" small PSU to go in a separate box with the main power transformer for the active circuit of the preamp, and have the 5V IC regulator at the PIC board as I supose its provided. Haven't look to the layout yet. |
|
|
| apassgear |
Yes, i see the regulator on the schematic so it will be on the PCB I guess.
Another question Brian. Should we tie both grounds?
I mean the PIC board ground to the star ground of the preamp?
Or this will contaminate the ground of the preamp? |
|
|
| BrianGT |
| quote: | Originally posted by apassgear
Yes, i see the regulator on the schematic so it will be on the PCB I guess.
Another question Brian. Should we tie both grounds?
I mean the PIC board ground to the star ground of the preamp?
Or this will contaminate the ground of the preamp? |
Not sure about this. I haven't given much thought to it. I am considering making a small power supply with one of those talema pcb mount toroids on it, on a seperate pcb for my power supply, and make it a regulated 5v. The vfd draws a lot of power, and worry about the long term reliability of the voltage regulator, as it is rated for 1A and it gets pretty damn hot, we had to upgrade the heatsink (originally had none). I am thinking about just making a 5v input to the passive volume control board, and then an input for the relay voltage, and if you are using 5v relays, just tie this the same. My roommate is doing almost all of the work for this project now, and I am trying to learn it as quick as I can so that I can help out with it.
What is a good plan for a regulated 5v power supply? It doesn't seem to be too important, as it only powers the control logic.
As for an earlier comment about the vfds making a lot of noise, if you look at the Passlabs amplifiers, all of the expensive ones have the vfd in the same chassis as the analog circuit, so it can't be that bad:
--
Brian |
|
|
| apassgear |
Good pic Brian, never saw that before. Its the Aleph P 1.7?
I think a good alternative for a reg PSU would be a similar to the one used by Nelson on preamps (and even Zen amp), series FET with a zener reference, but I don't know if the voltage is to low for this arangement. |
|
|
| BrianGT |
| quote: | Originally posted by apassgear
Good pic Brian, never saw that before. Its the Aleph P 1.7?
I think a good alternative for a reg PSU would be a similar to the one used by Nelson on preamps (and even Zen amp), series FET with a zener reference, but I don't know if the voltage is to low for this arangement. |
It is the Pass Labs X1 preamp:
http://www.passlabs.com/prodlit/x1lit.htm
The power supply doesn't need to be overly complex, since it is jus for control logic.
--
Brian |
|
|
| fcel |
Brian,
You mentioned that it's a passive volume controller and it appears that it's connected to the line out. Do I have to remove the two 2 gang pots that I have already installed on my BOSOZ? I know it sounds elementary but I'm asking anyway. |
|
|
| BrianGT |
| quote: | Originally posted by fcel
Brian,
You mentioned that it's a passive volume controller and it appears that it's connected to the line out. Do I have to remove the two 2 gang pots that I have already installed on my BOSOZ? I know it sounds elementary but I'm asking anyway. |
I am planning on connecting it directly to my BOSOZ with no pots whatsoever. I don't think that you will need the pots in the bosoz if you use the volume control that I am making. By passive, I meant that there are no active elements in the signal path.
--
Brian |
|
|
| moses |
| BrianGT, not too related, but why did your class choose to use a MAX232A verses a MAX233(A)? |
|
|
| fcel |
| o.k. then, I'll be looking forward to your completed design. |
|
|
| roddyama |
Brian,
It looks pretty interesting. I think I would be in for 2 or 4 channels if it didn’t happen too soon (too many irons in the fire right now). I know how you work though, it’s a glimmer in your eye one day, and the next day I’m looking at the finished product on your web site.:idea: :smash: :up: :note: :yes: So I’m wondering what your expected time frame might be for putting an order together?
Rodd Yamashita |
|
|
| BrianGT |
| quote: | Originally posted by moses
BrianGT, not too related, but why did your class choose to use a MAX232A verses a MAX233(A)? |
I would guess that it is purely cost. From digikey, the cost is:
max232a: $5.51
max233a: $10.76
They appear to do the same thing. The only difference would be that you need 4 external caps for the 232a, and the caps are cheaper then buying the better chip...
Is there any other difference?
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by roddyama
Brian,
It looks pretty interesting. I think I would be in for 2 or 4 channels if it didn’t happen too soon (too many irons in the fire right now). I know how you work though, it’s a glimmer in your eye one day, and the next day I’m looking at the finished product on your web site.:idea: :smash: :up: :note: :yes: So I’m wondering what your expected time frame might be for putting an order together?
Rodd Yamashita |
It won't be too soon, at least a month until I get the boards done for my next prototype, then more time to order real boards for the order. If you haven't seen the current pics, there are a few more on the website:
http://brian.darg.net/relay
My roommate actually put this thing together mostly in 2 days (you really have to love those end of semester deadlines and college procrastination... I do it a lot myself) He was originally going to do the PGA2310 for it, but it just wasn't clean, and it wasn't working out the way that he wanted to. I showed him the P1.7 manual, and he decided that it would be much cleaner, and quite easy to implement. The hardest part was writing the remote control code, which initally had sync. issues. He is rewriting it now to make it a universal learning remote control receiver, so that you can use any remote for the volume control.
He spent much more time on his earlier project, which he finished on monday, his USB alarm clock:
http://brian.darg.net/USB-Alarm-Project
It came out really impressive, and worked great. It had bi-directional communication over usb, and you could set alarms on either the computer, or the microcontroller, and you could pick the mp3 that you wanted the computer to play when the alarm goes off. Look at the pics, I thought it worked out great. He asked me earlier if I could think of a way to implememt usb into the volume control :D ...
I am slowly learning how to use the microcontroller, and am taking the class that he did next semester. The teacher put together a great book and board for the class, that anyone can buy and learn how to use fairly easily if you have basic programming knowledge. It has a C macroassembler, so you don't have to do straight assembly, although my roommate does anyway.
--
Brian |
|
|
| BrianGT |
I talked to my roommate some more about this project, and we have come up with a better plan for making this thing work:
-seperate boards for microcontroller and volume control board
-each volume control board is for 2 balanced channels. if you are not running balanced, simply only use one side of the relays.
-interfacing volume control boards with microcontroller with i2c, so that you can simply cable up to 16 volume control boards to this design, allowing you to have up to a 32 channel system.
-using the i2c addressing, the microcontroller can dynamically determine how many channels are present on the system, and they can all be configured seperately.
-each channel will have a label, and an individual setting (offset from master volume control)
-you will be able to change the balance of the board (mainly for 2 channel usage) and this will simply change the individual volume setting
-learning remote control receiver on the microcontroller board
-detailed menuing system for setup
-allow for user defined relay voltage, this would allow the user to pick whatever DPDT relays that fit the spacings provided (what spacings do relays have, are they standard?)
-microcontroller board pretty small
-volume control having short, clean signal path
That is it for now. We have the basic idea on paper how the board will be layed out. The microcontroller board will be quite small, and the relay board will be as small as possible, maintaining the shortest signal part possible (keeping it clean).
Mainly, this is intended to be a scalable design. Another possible advandage is that this might also allow for users to use other microcontroller boards, should you prefer to use an Atmel based board. The interface will be i2c, which is pretty common these days.
--
Brian |
|
|
| roddyama |
Brian,
How about extra outputs from the processor for relay selection of pre-amp inputs, or dual pre-amp outputs?:) Sorry if I'm complicating things.:ashamed:
Rodd Yamashita |
|
|
| moses |
| BrianGT: No, just the lack of external caps I find is usually a blessing(and frees up a little bit of PCB space), perhaps you guys are a little bit more proficient with the soldering iron :). The non-A parts are also cheaper, and just a tiny bit slower, which probably doesn't have much effect for the amount of data you're transfering back and forth. |
|
|
| BrianGT |
| quote: | Originally posted by roddyama
Brian,
How about extra outputs from the processor for relay selection of pre-amp inputs, or dual pre-amp outputs?:) Sorry if I'm complicating things.:ashamed:
Rodd Yamashita |
The microcontroller board will be capable of supporting this. Another board would have to be made, and if it is made using the i2c bus, it would be really easy to integrate this into the amplifier. It would just take running 4 control wires from the bus.
It doesn't really complicate things, but it would require making an extra pcb. The i2c chip for each board costs about $4 each in quantity, and one per expansion board is required. I don't know too much how i2c, but my roommate has a good grasp on how it all works (just finished two projects with it) and says that it simplifies the wiring a whole lot. each board would need vcc, ground and 2 control wires. It is a simple 2 wire serial bus developed by Phillips:
http://www.semiconductors.philips.com/buses/i2c/
This definately is not my first priority for this, but once the basic 2 board solution is worked out, this will be next on the list.
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by moses
BrianGT: No, just the lack of external caps I find is usually a blessing(and frees up a little bit of PCB space), perhaps you guys are a little bit more proficient with the soldering iron :). The non-A parts are also cheaper, and just a tiny bit slower, which probably doesn't have much effect for the amount of data you're transfering back and forth. |
Soldiering wasn't too bad for this. The volume control won't have this chip on it, since it won't be using the QuikBug (c) software. The chips will just be programmed with a programmer and put on the board.
As far as data transfers, it can take quite a bit to program this thing (1 minute) for large programs. It is quite a large pic. My roommate actually wishes that it went much faster.
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by csd
Also interesting to note is that it isn't obvious how to switch the relays - it's not just a simple binary count. It took me a while to figure it out. For the first half of the range, you count up from zero in steps of two. For the second half of the range, you count down from 255 in steps of two.
Craig |
I am quite confused exactly what you mean by this. I can't seem to picture why you say that it is this way... can you elaborate a little more on this?
Another possibility might be using a table of values, and referencing that, and finding what values closely relate to the db scale, and scale it that way. A couple of relays could be added to the end to give 1024 total steps, then from this, a db scale could be determined as closely as possible, and put into a table. Is this possible through modifying the current configuration.
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by roddyama
Brian,
How about extra outputs from the processor for relay selection of pre-amp inputs, or dual pre-amp outputs?:) Sorry if I'm complicating things.:ashamed:
Rodd Yamashita |
Hell... since it is i2c, you could even run i2c temperature sensors from the processor board to each one of your monoblocks and monitor temperature from the display. I know that I am for mine. They don't cost that much either. There is a temperature on the development board, and I still like to use it to see what temperature my room is.
--
Brian |
|
|
| BrianGT |
I have done some research into the vfds, here is the pricing direct from Noritake:
Noritake 7000 series
http://www.noritake-elec.com/7000.htm
http://www.diyaudio.com/forums/attachment.php?s=&postid=94665
GU140X32F-7002 (one in my picture above, beautiful display)
1-9 $96
10-49 $88.80
50-99 $72.10
100-199 $54.80
GU140X16J-7002 (half the height of the above vfd)
1-9 $61.80
10-49 $57.20
50-99 $46.40
100-199 $35.20
GU280X16G-7002 (twice the width of the above vfd)
$96.90
$89.60
$72.70
$55.20
In terms of rotaty encoders, I am thinking of using a Greyhill 32 position with pushbutton:
grayhill pn: 62A11-01-050S
digikey pn: GH3048-ND
Datasheet: http://38.254.60.87/embrace/IMAGES/PDF/bltn725.pdf
Would anyone be interested in putting together a group buy of vfds for a project like this?
--
Brian |
|
|
| Variac |
Brian says this display is super cool. Since he has seen it and worked with it I think we should definately get one of the versions of it.
The fact that it can display mutiple data in separate frames sounds perfect for our needs. It is likely we will often want to display different text at he same time. Maybe the text gets big while adjusting a function, then recedes to a small window after.
OK there are a million options. In fact I can see Brian or others releasing later versions of the PIC with more and more groovy features. In fact I need that temp function to monitor how much I can turn up the variac on my Son o Zen. Maybe the choice of the size he shows or half size should wait to see the number orderd. If a large quantity is ordered the bigger display is reallly not that much more.
There is an amazing price break at 100 units, and I think we might get that many orders. The biggest problem is the logistics I think. That would be 1800 relays for instance. Someone would certainly have to help Brian finance the thing, unless he decides to go into business. Imagine the great price we would get on almost 2000 relays!!
If the basic 2 channel w/ display setup has all the features Brian mentioned above, plus as many more as the human brain can devise, then it would be a screaming bargain at $150. (just a guess at the 100 order price). Still, we're talkin' $15000 dollars here- how can this be done?
I'm definetly in for a stereo version! |
|
|
| dorkus |
| i'm not sure the cost savings on 100 units is worth the headache in logistics... |
|
|
| Variac |
Dorkus
On a lot of the parts (relays) I think I will have to agree with you,
But the display is HALF PRICE if you get 100!!! |
|
|
| JordanG |
Hey, so I'm glad to see such enthusiasm in this project. I’m Brian’s roommate btw.
Anyhow, I was wondering about the resistor-relay network. When I made my initial circuit, I used a handful of the resistor values from the volume circuit in the Aleph P. However, Craig made the point that specifying the resistor network was a tricky task; never mind deciding on how to trigger the relays. Should we be attempting to specify a newer, more precise resistor network, or would it be sufficient to use the Aleph P version? I suppose we could use more than 8 relays per balanced channel; however I’m not really sure how to approach designing a new resistor network. Anyone have any ideas where I might to go to read up on the subject? Also, how is the tradeoff between the size and cost of the relays vs. the precision that can be obtained? For instance, it would certainly be nice to have .5dB steps from -70 to 0, but I don’t know how realistic it would be.
Any insight would be appreciated.
Thanks
--Jordan |
|
|
| BrianGT |
| quote: | Originally posted by JordanG
Hey, so I'm glad to see such enthusiasm in this project. I’m Brian’s roommate btw.
|
Indeed, this is the guy that makes it all happen.
--
Brian |
|
|
| Variac |
Hey Jordan,
Brian is almost a folk hero around here. If you pull this off you will be a folk LEGEND! I guess the resistor conundrum can be worked on while you design all the other factors?
Obviously we need to decide these issues before the board layout is designed:
how many relays? Can one or two more give precise 1 db increments? Do we really care?
Is the shunt design really better? if so, how is it implemented?
Later on we can figure out how to deal with the huge cost of this.
At the minimum Brian would need to supply the pic and PCB. |
|
|
| JordanG |
| quote: | Originally posted by Variac
Is the shunt design really better? if so, how is it implemented?
Later on we can figure out how to deal with the huge cost of this.
At the minimum Brian would need to supply the pic and PCB. |
Shunt design? Could you point me in the direction of such a design, as I'm not familiar with it. Truth be told, I'm a humble computer engineer.. Almost no analog knowledge whatsoever (that's what Brian's for).
And speaking of what is to be supplied, I'd like to say that the design I'm working out with Brian is starting to look pretty nice. I'm envisioning a control board with a microcontroller at its heart (most likely a PIC18 series). The board would initially require that nifty Noritake display (gotta start somewhere), though I've been considering lower cost displays after we get the intial design down. The passive volume circuit boards would each support 2 balanced channels, and up to 5 boards could be tied to the control logic (for up to 10 channels of balanced audio per control board). The control circuit will be setup to "sense" the presence of additional volume boards, so you're in no way tied down to 2 channels if that's what you start with. The "sensing" feature can be added at almost no additional cost to the end user, as it is a function of the control circuitry that I've chosen to use (I2C). A learning remote feature is also forthcoming, so that the user could use any decently modern IR remote. Also, the PIC18 has an EEPROM on it, allowing allowing us to do things like setting the left and right channels, naming different channels, saving balance settings, brightness settings, etc.
Of course, anything can change at this point, that's just the current status.
--Jordan |
|
|
| harvardian |
Hi Jordan, Brian,
A friend and I are working on a PIC18 project in which one of the boards is a USB-CAN converter. We wrote a USB bootloader. The code downloads in a couple of seconds. We also have a board that does a servo/stepper with a CAN interface. We also have a bootloader for the CAN board in which we send the code from the PC to the USB board and then on to the CAN based board. For a 28K program (almost 80% of 18F458), the sownload takes only 6 seconds.
Dale
P.S. I am using the CCS C compiler. A lot easier than assembly. |
|
|
| roddyama |
| quote: | Originally posted by JordanG
Shunt design? Could you point me in the direction of such a design, as I'm not familiar with it. Truth be told, I'm a humble computer engineer.. Almost no analog knowledge whatsoever (that's what Brian's for).
|
Hi Jordan,
Take a look at this thread started by HH/Fred Dieckmann (Brian can fill you in on that story). He presents a passive shunt-type volume control with no resistors in the signal path. It truly is a minimalist approach. He uses a pot for his circuit but that can just as easily be discrete steps.
http://www.diyaudio.com/forums/show...ve+volume+shunt
Brian,
I know that Dorkus mentioned this on page one of this thread. Sorry to be a pest, but have you considered the shunt approach?
Rodd Yamashita |
|
|
| gnomus |
I guess I am a regular customer now as I bought Brian's really nice Aleph boards. I would be interested in this project. I was thinking about volume and selector controls for XBOZ . And I do not know if it would work well but as a passive control in an AlephX intergrated.
It is not clear to me how many single ended and balanced inputs and outputs will be available.
I am interested in one pair of balanced inputs. For a DAC. Is this planned? |
|
|
| BrianGT |
| quote: | Originally posted by roddyama
Brian,
I know that Dorkus mentioned this on page one of this thread. Sorry to be a pest, but have you considered the shunt approach?
Rodd Yamashita |
Well, you have to start somewhere, so the current plan for Jordan and I is to get prototype pcbs done for the 8 relays and see how it works out. While the boards are getting made, we will also work on designing some other boards based on a different methodology, such as a shunt volume control. I want to try the P1.7 version out first and see how it actually works in practice, the built something else to compare it to.
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by harvardian
Hi Jordan, Brian,
A friend and I are working on a PIC18 project in which one of the boards is a USB-CAN converter. We wrote a USB bootloader. The code downloads in a couple of seconds. We also have a board that does a servo/stepper with a CAN interface. We also have a bootloader for the CAN board in which we send the code from the PC to the USB board and then on to the CAN based board. For a 28K program (almost 80% of 18F458), the sownload takes only 6 seconds.
Dale
P.S. I am using the CCS C compiler. A lot easier than assembly. |
Dale,
That sounds like quite an interesting and complex project. Do you have any sort of webpage with more details on this? Jordan has tried the CCS C compiler, but said that it made timing issues with stuff like IR communication more difficult, due to the stack management.
What parts do you use for the above mentioned project?
--
Brian |
|
|
| BrianGT |
| quote: | Originally posted by gnomus
I guess I am a regular customer now as I bought Brian's really nice Aleph boards. I would be interested in this project. I was thinking about volume and selector controls for XBOZ . And I do not know if it would work well but as a passive control in an AlephX intergrated.
It is not clear to me how many single ended and balanced inputs and outputs will be available.
I am interested in one pair of balanced inputs. For a DAC. Is this planned? |
The design will be modular, so there will be one control board, one volume control board for each 2 balanced channels, and there is a possibility of an input relay selection board also.
For now, we are just working out a two board solution with one control board, and one volume control board. Jordan's post above explains this a bit better.
My plan for this is to use it attached to the balanced output of my bosoz, then going into my aleph-x. This is my plans for it, but it can be used many other ways. You could even run 4 unbalanced channels through one 2 channel balanced board if you wanted. I am going to use another board to run between my unbalanced cd player and my amplifier also.
--
Brian |
|
|
| csd |
Oops...
There was an error in the program I wrote to graph the attenuation vs. step, so please ignore that graph and the conclusions I derived from it!
The attenuation is a linear function of step, similar to a linear potentiometer. In order for the function to be smooth, each resistor must be twice the resistance of the previous, and the P1.7 design is close to this. Of course, what we actually want is a logarithmic response, so this isn't ideal. What this design does give us is a minimum of relays.
We could add a fixed shunt resistor, as in Rod Elliot's notes here:
http://sound.westhost.com/project01.htm
This would make the response closer to logarithmic.
Using 10 relays provides up to 60dB of attenuation, which some people might like. I personally think 48dB with 8 relays isn't enough.
Craig |
|
|
| fcel |
Brian/Jordan,
I was wondering if the systems that you guys are designing would solve the problem of turn on/off thumb from the BOSOZ. (See this thread for some of the ways that I have already tried http://www.diyaudio.com/forums/show...on&pagenumber=7)
From what I have read so far on this thread, it appears that the relays should be able to short out the output at turn on/off to supress the turn on/off transient. I would like to know if you guys can implement this into the design if you have not considered this option yet. |
|
|
| Havoc |
Sorry to fall in late, but I have been away for a couple of days.
Why not a constant impedance attenuator? I made one (stereo) with 6 relays in each side. With a 1dB step you get 63dB max attenuation and a nice log law. You can make these balanced or not, it only determines the relay you need. Mine has a 1k impedance, but you can go for whatever you want. |
|
|
| Variac |
Havoc:
do you have a reference or scematic of this for the ignorant (me)among us ? |
|
|
| roddyama |
Havoc,| quote: | Originally posted by Variac
Havoc:
do you have a reference or scematic of this for the ignorant (me)among us ? |
"Me" too.:)
Rodd Yamashita |
|
|
| BrianGT |
| quote: | Originally posted by Havoc
Sorry to fall in late, but I have been away for a couple of days.
Why not a constant impedance attenuator? I made one (stereo) with 6 relays in each side. With a 1dB step you get 63dB max attenuation and a nice log law. You can make these balanced or not, it only determines the relay you need. Mine has a 1k impedance, but you can go for whatever you want. |
I am interested to see what you are talking about, can you post a schematic, or guide me to somewhere where I can learn about this?
--
Brian |
|
|
| hifiZen |
Hmm... this looks like an interesting project. I've tried a couple of times to come up with a reasonable relay and resistor-network based volume control, but I've never been satisfied with the results since I've always been after a true logarithmic curve, or something very close. I really want to have consistent steps in volume from one end of the scale to the other, and I've always been super-annoyed by volume controls that couldn't get that "in-between" setting, so fine steps are important to me. Otherwise, one position is just a bit too quiet, and the next is just a bit too loud. :irked: Combined with the prohibitive cost of relays, and the crappy log approximations one could botain with such networks, I've been increasingly drawn towards devices like the PGA2310. But, a group order could be just the thing to bring the project to maturity at an affordable price. :up:
I think you've got the right idea to separate the microcontroller from the attenuators, as some people may want to do their own microcontroller or integrate the code into a controller doing other things as well (like me for instance), and/or put the micro in a separate chassis. But, I'm puzzled why you decided to put four banks on one attenuator board. Why not do 2 banks per board? Then you can use a single board for either one balanced channel, or a stereo pair...
Next, can I suggest something simpler than I2C for the interface? If I2C is used, then it would have to be a proper (read: meets the spec) implementation so that someone who connects these to an I2C bus with other devices on it can still use the boards. But, a proper I2C implementation would require a micro right on the attenuator PCB, which isn't an optimal solution, since the purpose of a separate micro is to get it and it's noise away from the analogue signals. I would use simple SIPO shift registers to control the relays. This would only require a 2-wire interface, and can be bit-bashed and operate at whatever frequency you want, unlike I2C which is fixed at 100 or 400kHz. Since relays are so slow to react to a change in drive signal, you can easily shift in the entire control binary value for many channels before the relays have any chance to react. Plus, you can daisy-chain as many SIPO shift reg's as you want, so your total number of channels is almost unlimited...
Jordan: the shunt method works like this: you basically have a fixed series resistor connecting to an array of switched resistors to ground (the "shunt" elements). With all the resistors turned off, your attenuation is zero, and as you shunt more resistors to ground, the attenuation increases. The shunt method reduces sensetivity to non-ideal characteristics of the shunt devices. I believe the most recent incarnation of attenuator at Pass Labs is actually using FETs as shunt switches... Which brings me to another point: FETs are way cheaper than relays, which makes it affordable to use more switched elements, thus acheiving greater precision. Drive requirements are simple too, since they can typically be turned on directly with logic-level signals, so theres another cost savings (a relay requires a drive transistor and a diode to prevent switch-off voltage spikes from destroying the transistor). It would take a bit of work to figure out the optimal arrangement, but I'm sure an arrangment of vaious combinations of parallel shunt elements could be found which gives pretty precise 0.5dB steps... not necessarily a binary progression, but more precise, and following a true log curve, thus requiring more shunt elements than a simple binary progression. The cost saving from switching to a FET-based shunt scheme would enable this.
Havoc: please do provide more info on your attenuator! Preferably a schematic and the necessary formulae to calculate resitor values.
Group buy on Noritake VFDs sounds like an excellent idea. You can definitely count me in! These graphics displays are hard to get at a reasonable price. I've seen the Noritake units too, and they are <i>super</i> nice VFDs! As nice as they come... The logistics of 100 orders isn't actually so bad, and well worth it for the price IMHO.
BTW Brian, Jordan - you know that the PGA2310 needs a low impedance signal source (<600ohm), right? I wonder if that mp3 player I see in your photos meets that requirement... I've heard that these type of device are very well regarded. But alas, the PGA3210s sitting in my parts bin sit idle while I deal with other things. Hopefully I will get a chance to rig them up and have a listen before long... |
|
|
| kuribo |
| I have been using a transformer volume control, and it beat hands down the highly regarded relay based, resistor attenuator made by Placette. I wonder if you have considered using relays to switch the secondary taps of these transformers to adjust volume? I suppose you would need a double relay for each set of taps, which would mean around 24 relays....Currently I am using a high quality rotary switch made by Seiden, but wonder if high quality relays might be better....See www.bentaudio.com for info on the transformer volume control.... |
|
|
| roddyama |
Chad,
So you crawled out from under your mass of board orders did you?:tons: :Popworm: Who said you could come out and take a break (Just kidding;) ).| quote: | Originally posted by hifiZen
...The shunt method reduces sensetivity to non-ideal characteristics of the shunt devices. I believe the most recent incarnation of attenuator at Pass Labs is actually using FETs as shunt switches... Which brings me to another point: FETs are way cheaper than relays, which makes it affordable to use more switched elements, thus acheiving greater precision. Drive requirements are simple too, since they can typically be turned on directly with logic-level signals, so theres another cost savings (a relay requires a drive transistor and a diode to prevent switch-off voltage spikes from destroying the transistor). It would take a bit of work to figure out the optimal arrangement, but I'm sure an arrangment of vaious combinations of parallel shunt elements could be found which gives pretty precise 0.5dB steps... not necessarily a binary progression, but more precise, and following a true log curve, thus requiring more shunt elements than a simple binary progression. The cost saving from switching to a FET-based shunt scheme would enable this.... |
Excellent.
Rodd Yamashita |
|
|
| JordanG |
| quote: | Originally posted by hifiZen
But, I'm puzzled why you decided to put four banks on one attenuator board. Why not do 2 banks per board? Then you can use a single board for either one balanced channel, or a stereo pair... |
<shrug> Compromise between modularity and affordability I suppose. More pcb's require additional cost, right?| quote: | | Next, can I suggest something simpler than I2C for the interface? ... a proper I2C implementation would require a micro right on the attenuator PCB |
Yeah, I know. I2C might be overkill, but its easy to work with, and it makes for an instant modular design. Noise was an issue I hadn't considered, but PCB design is able to handle that issue, isn't it? But the main reason I use I2C? TI ships 8 8-bit I2C IO expanders per sample request, so I always prototype with them anyhow. I'll definately take a simpler serial interface into consideration however.| quote: | FETs are way cheaper ... Drive requirements are simple too ... FET-based shunt scheme would enable this
|
I'm in no way bound by the relay based volume scheme, its just one that was easy to understand and implement with limited knowledge of analog circuitry. If I knew how to do a FET-based shunt design, and it would satisfy audio enthusiasts, I would use it in the project. However, I will say that ultimately I will make certain compromises in order to keep the design tractable. For instance, the FET-based shunt design might be a second phase volume board. Ultimately, I want to get a design out there for people to start playing with. The first version won't be the best version, but it will be a good starting point.| quote: | a relay requires a drive transistor and a diode
|
Here is where my lack of analog knowledge really starts to shine. So I have a few questions for you. Do you still need the diode if the relay coil isn't polarized? Also, what does the drive transistor get you? In my prototype design, I used 5V relays that were driven by flipflops. From what you said, I doubt that was sufficient. Feel free to enlighten me. :)| quote: | | you know that the PGA2310 needs a low impedance signal source (<600ohm), right? I wonder if that mp3 player I see in your photos meets that requirement... |
I doubt the mp3 player met the requirment, but then again, I didn't run the mp3 player through the PGA2310 circuit. It was just in the picture because Brian thought it would look nice next to my other circuit. The circuit I used was based on Mark Hennessy's design.
http://www.mhennessy.f9.co.uk/preamp/main_audio.gif
Thanks for all the commentary though. You brought up some very good points
--Jordan |
|
|
| JordanG |
Ok, question. I heard that digitally controlled pots weren't held in very high esteem. Is this true? The reason I ask: central to this design is the ability to change the volume with an IR remote. If the design includes a pot, then it needs to have some kind of hands-free control.
I mean, I suppose I could rig a stepper motor to it. But that would be an all together different project.
--Jordan |
|
|
| roddyama |
| quote: | Originally posted by JordanG
Ok, question. I heard that digitally controlled pots weren't held in very high esteem. Is this true? The reason I ask: central to this design is the ability to change the volume with an IR remote. If the design includes a pot, then it needs to have some kind of hands-free control.
I mean, I suppose I could rig a stepper motor to it. But that would be an all together different project.
--Jordan |
Hi Jordan,
Forgive me if I’m missing something (totally in the realm of possibilities), but wasn’t the plan to use an encoder for manual volume control input? I assumed also that the IR input would increment a digital counter. This would not require pots or stepper motors. The only specific problem I see with using encoders and counters for the volume input is the need for the “volume control” program to have a constant memory for last setting, and a visual volume setting indication.
Rodd Yamashita |
|
|
| Variac |
I think they have 2 banks of resistors on each board only, but because each one is DPDT then it can control a balanced input .
Of course it could also control 2 unbalanced lines but then the balance couldn't be adjusted. I think the compromise of 2 dpdt banks of relays on each board is a good one, also the display. |
|
|
| JordanG |
| quote: | Originally posted by roddyama
wasn’t the plan to use an encoder for manual volume control input? |
Well, the encoder just provides my PIC with information about the direction and speed the encoder is being turned. I then take that information and use it to modify a volume state variable. The IR remote will be modifying that same state. And saving the state variable isn't a problem, because the pic I'm using has an EEPROM on it.
--Jordan |
|
|
| Brett |
| quote: | Originally posted by kuribo
I have been using a transformer volume control, and it beat hands down the highly regarded relay based, resistor attenuator made by Placette. I wonder if you have considered using relays to switch the secondary taps of these transformers to adjust volume? I suppose you would need a double relay for each set of taps, which would mean around 24 relays....Currently I am using a high quality rotary switch made by Seiden, but wonder if high quality relays might be better....See www.bentaudio.com for info on the transformer volume control.... |
kuribo,
I use exactly the same setup an have been wondering about a remote control for volume, selection, mute and absolute phase (part of the source selector). I have a bunch of suitable relays but assume that there will need to be "timed" to ensure make-before-break action. A project for the near future. |
|
|
| Variac |
Suddenly I'm confused.
I had always assumed that Brian meant for this to be used as a control for a pre-amp, either soboz or sodork. In that case the impedence isn't as much an issue, right? . Do we have to decide whether this is a stand alone unit to be used instead of a preamp or a volume control FOR a preamp in order to decide on a design?
If so we should do one first, then the other later. |
|
|
| kuribo |
| The transformer volume control I use has a remote volume control which I put together using a radio remote controling an on/off relay for a small gearmotor. It works well, though not the optimum solution. I understand that John at Bent has put together a stepper motor remote that should be better than mine, but I can't help but wonder about an all relay system to ghet rid of the motor and switch. The make before break is something I didn't thing about....Hhhmmm, wonder how that can be solved.... |
|
|
| Brett |
| quote: | Originally posted by kuribo
The make before break is something I didn't thing about....Hhhmmm, wonder how that can be solved.... |
I presume a uprocessor would be needed to switch the new attenuator step or source in before the previous one was deselected and would need to account for the timing of the specific relay's ton and toff. Timing loops in software would be easy to do. I don't want thumps when switching. |
|
|
| jackinnj |
| and with a lot less effort. Below is an HP3488a switch -- these go for about $40 on the Bay with a card or two. (Originally they were about $3400). The 44470A relay card will switch 10 signals from up to 250 V, you can fit 4 cards in the switch. There is an analog card, a breadboard card, a video card, another relay card etc. It can be programmed over the GPIB/HPIB buss using a National Instruments card. You get a very high quality board, gold plated to the usual HP standard, excellent power supply, rack cabinet etc. They are inexpensive since they were used by companies like Cisco, Lucent and Nortel and a lot of this stuff is just disgorged onto the market. <p> Of course, the XYL factor may dictate that HPIB cables and PC's be kept out of the listening area, but this rarely stops DIY'rs. I daresay that you could probably fit an entire preamp within the space and provide all the controls with the 3488a (channel switching, volume, record, mute etc.) |
|
|
| tvi |
Try here for an example of a:- <a href="http://www.geocities.co.jp/Technopolis/5053/amp/t_att.htm">Fixed impedance type ATTENUATOR
<img src="http://www.geocities.co.jp/Technopolis/5053/amp/20kATT.gif"></a>
Its in Japanese, you may want to try <a href="http://www.worldlingo.com/wl/Translate">worldlingo.com</a> for a translation.
Some of the Values for R1 might prove hard to obtain.
Regards
James |
|
|
| jgwinner |
| quote: | | Combined with the prohibitive cost of relays, |
One thing I did recently was use AUTO relays ... I got some from Mouser for pretty cheap I needed 12v anyway (the xformer I have has a 12v secondary).
I was trying to get 30A relays for my OptiMOS home theater 2Kw power supply, and I was having a hard time finding them - so on a lark I got some $2 relays that'll handle that (probably were headlight or starter relays). They aren't physically very large, and I was trying to NOT get 'miniature' relays. I'm not sure how they sound yet, but I'm planning on using one relay per channel for the output relays. For the power supply soft start they are pefect of course.
I'm guessing here though, you need make before break, and I don't think these type of relays will work. |
|
|
| mhennessy |
I've been following this thread with interest, and there's just a couple of comments I'd like to add...
Is the PGA2310 that bad?
I know that it isn't going to be as transparent as a passive solution - this goes without saying. But, I'm wondering about how you came to decide that it is as unacceptable as you say... During a past discussion in this thread, I asked some questions about power supply decoupling and capacitor type selection, and made some suggestions that might improve the sound quality. There was no response at the time, but did you get the chance to try them?
Constant-impedance attenuators
Already mentioned here, but in my view, this is the only way to go if you want to use relays... I've considered it myself in the past (albeit using analogue switches rather than relays)
It's simple, in theory. You decide on a system impedance (say, 600 ohms). Then, you build a number of attenuators that have a 600 ohm input impedance, 600 ohm output impedance, and the attenuation required. Eg, 1dB, 2dB, 4dB, 8dB, 16dB, 32dB... See the pattern?
Next, relays switch them in and out of circuit. You'll need a dual-pole relay for each channel... Therefore, they are cascaded until you have the required attentuation - this part of the maths is simple, because 4dB plus 8dB gives you 12dB
You need these so-called constant-impedance attenuators so that the attenuation of one isn't altered by the loading of the next one...
The hard part is choosing the resistor values for each attenuator. But, this information might already exist somewhere on the Internet...
So, just my 2 cents worth...
Seasons greetings :) |
|
|
| Havoc |
| The idea is explained by mhennesy. You make a chain of attenuators that have a certain impedance when loaded by that impedance. As I do not have anything on my pc to make a drawing or scan right now, I will try to put something up tomorrow. |
|
|
| hifiZen |
| Ah ha! Finally a workable scheme. Thanks Mark and tvi! Are you taking notes, Brian? ;) |
|
|
| hifiZen |
| quote: | | So you crawled out from under your mass of board orders did you? |
Indeed I did. It got kinda stuffy after a while... ;)
JordanG:
Have a look at these parts:
Toshiba VHC164
Toshiba TD62083AF
Both ICs are available from DigiKey, and I've used them in concert before for a similar type of project. Indeed, logic gates are really not up to the task of reliably driving a relay coil, even if it's a pretty small one, so in general a driver transistor is needed to switch the coil currents. That's what the second IC is for. It's basically a bunch of OC (open collector) driver transistors in a compact package. I used it to drive LEDs, but relay coils work just the same.
The 74VHC164 is the SIPO (serial in parallel out) shift register device I used. All that's required is one data line and a clock to shift the data bits in. Quite simple, I think you'll agree, and much less hassle than a micro or I2C chip, although you may need to write a quick little software routine to do the bit-bashing. I don't think this particular chip can be daisy-chained, but you can see how easy it would be to control numerous banks of resistors with a bunch of daisy-chained SIPO registers... Plus, you can write data to this chip at 20HMz, or whatever speed your micro can muster, as opposed to the pokey 100/400kHz clock of I2C.
Regarding relay control, in general it does not matter if the coil is polarized. Because it's a coil, it is an inductive element, and that's all that matters. Shorting one end to ground with the other end tied to the supply voltage will cause a current through the inductor, hence storing energy in it's magnetic field. If you switch it off in a hurry, the inductance will try and force a current to flow, in the process generating whatever voltage it needs to dump this stored energy. This usually would result in overvoltage and destruction of your switching device, or maybe an arc inside a mechanical relay which would cause premature aging and failure. I suspect that your logic gates driving the relays are having their ESD protection paths exercised each time a relay turns off. All that's required to solve the problem is a diode in parallel with the coil. In normal operation, the diode is reverse biased, but when the switch turns off, the diode becomes forward biased, carrying the current being dumped by the coil, and thus preventing any excessive voltages from forming. Incidentally, many car ignition systems work the same way. A large coil is shorted between +12V and ground, building up a big current in the coil. Then, the coil is disconnected from ground, and the floating terminal jumps up to extremely high voltage in an effort to dump the coil's energy. At some point, it reaches a high enough voltage to arc across the spark plug's gap, and the coil's stored energy discharges through the ionized gas. |
|
|
| janneman |
| quote: | Originally posted by mhennessy
[snip]... I've considered it myself in the past (albeit using analogue switches rather than relays)[snip] |
Mark, can you tell me which switches you used, what were your impressions sound-quality wise?
Thanks,
Jan Didden |
|
|
| ftorres |
Hi there.
Adding to the confusion, I'd like to drop my two cents here. Two or three years ago, I made a relay-based volume control for my BoSoZ. It's driven by a uC (AVR), and has a "make before break" feature. For people interested, I posted the schematic here on DIYaudio, along with some infos. It's working faultlessly and more than fine. If you need more info, let me know. |
|
|
| JordanG |
| quote: | Originally posted by hifiZen
The 74VHC164 is the SIPO (serial in parallel out) shift register device I used. All that's required is one data line and a clock to shift the data bits in. Quite simple, I think you'll agree, and much less hassle than a micro or I2C chip, although you may need to write a quick little software routine to do the bit-bashing. I don't think this particular chip can be daisy-chained, but you can see how easy it would be to control numerous banks of resistors with a bunch of daisy-chained SIPO registers... Plus, you can write data to this chip at 20HMz, or whatever speed your micro can muster, as opposed to the pokey 100/400kHz clock of I2C. |
Well, if it were just a speed issue, then the bit-banging would get me a data rate of about 2.5MHz (internal clock frequency of the PIC I'm using). And that's only going to be 5-6 times faster than I2C (400kHz), max. In practice, bit-banging is going to be maybe 4-5 times faster. And since we're dealing with micro seconds, 4-5 times faster shouldn't be noticable. But I'm also interested in making a conviniently modular solution. And with I2C, I can use the PIC to determine how many passive volume boards are connected. So if you originally had the pic and a single volume board (2 channels), and you wanted to add another two channels, you simply tie an addional volume board to the I2C bus lines, send some kind of EEPROM reset condition to the PIC (through some menu option or jumper setting), and the PIC will recognize the new board and act as if there were always 2 boards (4 channels). Now, I2C isn't a hassel on the PIC side because Microchip built an I2C module into the chip, in order to make I2C more convinient. Price isn't a major issue, as the 16-bit I2C I/O expander I'm looking at is around $2.20 a part (Phillips PCA9555). And the only other thing I have to worry about is the noise from the chip...
You do make really good points for using that SIPO chip though. The best thing I can tell you is that after I get the I2C volume board prototyped, I'll take another look at bit-banged serial.
Anyhow, about the noise concern. The PCA9555 is a 24-pin SOIC part. Would it be sufficient to surround the chip with ground plane, or is there a better way to deal with the noise?
--Jordan |
|
|
| JordanG |
| quote: | Originally posted by hifiZen
Regarding relay control ... |
Many thanks for the information regarding the drivers and diodes. It was exactly the information I was looking for.
--Jordan |
|
|
| mhennessy |
| quote: | Mark, can you tell me which switches you used, what were your impressions sound-quality wise?
|
Hi Jan,
This wasn't really for a mega-serious hifi application - it was for my NICAM tuner - this used standard 4052 CMOS switches to select between NICAM and FM, and do the muting. As there was two spare buttons on the remote, I was planning to have a simple volume control - maybe 2dB steps, and using a simple 4-bit counter that would give 32dB of total range. My hifi wasn't remote-controlled, so this would have given me some control when watching TV...
I never got around to implementing it for a number of reasons - mainly because of the poorly-defined (and rather variable) RDS(ON) of the CMOS switches. That wasn't a problem for the switching mentioned above because they were followed by a buffer with a high input impedance. However, for a constant-impedence attenuator, the resistance variations would cause inacuracy and significant distortion...;(
You could get around that by using high-values resistors, but you'd introduce Johnson noise....
RDS(ON) varies with common-mode voltage
Cheers,
Mark |
|
|
| mhennessy |
I know that this thread is discussing passive relay control, but I've just had an idea... !:idea:
Using an inverting stage, you could make a mix-amp. Each input has a different gain weighting, and by choosing the resistor values carefully, you can make them dB's in powers of two, as shown. This isn't too difficult in practice...
Then, addres the relays using normal binary... It's a crazy plan, but it might work |
|
|
| janneman |
Mark,
Thanks. You know of course that there are modern switches with <5 Ohms Ron with very flat resistance. They would be OK used in a virtual ground mode.
You last scheme might work, but note that in the -64dB pos the resistance ratios around the opamp are > 1: 1000. Not easy, you know, noise and all that. An R/2R ladder (available as SIPs from Vishay) with a couple of the above mentioned switched could do the trick.
Jan Didden |
|
|
| mhennessy |
Hi Jan,
| quote: | | An R/2R ladder (available as SIPs from Vishay) with a couple of the above mentioned switched could do the trick. |
But the steps would be linear!
;) |
|
|
| sonnya |
| It would require a lookup table in a micro to make them log. |
|
|
| phase_accurate |
sonnya's method is one possible solution, and quite an elegant one at that.
The other soution is not bothering at all about logarithmic behaviour. While I admit that there are situations where it is important to have logarithmic readout, it definitely isn't, at least not to me, for the volume setting of an amp.
I usually set the volume so that it is pleasing to my EARs and not to make any kind of readout pleasing to my eyes. ;)
Regards
Charles |
|
|
| janneman |
| quote: | Originally posted by sonnya
It would require a lookup table in a micro to make them log. |
Right. AD had some MDACs that were really 17-bit R/2R ladder DACs with internal conversion through a look-up table so the input dig code would be in equal steps of I believe .375dB for 70dB range. AD7745 or something. Discontinued a couple of years ago because you guys insisted on buying stepper switches ...:rolleyes:
Jan Didden |
|
|
| guido |
Hi,
Made a preamp (tube one) with:
- PIC 16F84 controller is in sleepmode when no code is received.
- I2C IO expanders used to control the relays for switching the inputs, volume and power of the preamp.
- Bourns position decoder (ACE-128) on the volume, so it's position is know at all times.
- The conversion table for the position decoder is stored in an I2C EEprom. (24C02)
- The PIC can be programmed in-circuit, a DB9 connector at the back of the preamp can be connected to a serial port.
- Controls my Kenwood tuner using the Kenwood 8-bit XS protocol (on/off, change presets)
Using shunt-volume regulation at the moment with Alps pot.
My relais look like the pass X1 relays used for input selection...!
I am using SDS relays (filled with some gas, sliding contacts if i remember correctly)
Anyway, If you need some routines for the pic (the 16f84 is much smaller!!), let me know.
Click on my www below for more info. Don't use a display on purpose. Displays have their own controller, clock, noise...
Relays are driven with ULN2004, Alps and one relay with TC4469.
IO expanders are 8-bit PCF8574. If you need more info on this, let me know aswell.
All is working very well, no problems so far (two of these controllerboards are in use).
Greetings,
Guido |
|
|
| Havoc |
I took a bit longer than expected.
There is an unbalanced and a balanced one. Each only have 1 stage, but you can add as much as you like (well, a 120dB attenuator may not be useful). I put in a L-type attenuator, but you an as well use T, pi or bridged T (like already showed in a post here) for the unbalanced. The balanced is a H-type, but O works just as well. Formulas for all those can be found on the web or in HF textbooks.
The resistor at the output buffer is needed only once, but is needed. This is the impedance you design for.
The one I made is of the L type with 6 stages. Impedance 1k and steered by a simple up/down counter. |
|
|
| Helix |
I thought i might drag this thread up again, it has gone a bit quite.......
Yesterday i had a day to spare so, i thought i might have a go at making one of these attenuators ala unblanced part of pass pre-amp p1.7 (i think, u know the one with the 8 relays and res from 8k to 60ohm) (also i hadn't read this thread up to this point). So i grabbed some relays from the local shop and fired up my Elecktor 8052 flash project board and shuved in some code. I finally finished wiring up the bread board (prototype board) and got it working. Input to the relays was just BCD with 00 being all relays off and FF being all on.
I connected a 12 v DC input and put my multimeter and saw the output going up and down nice with clicked precision. However, that is when it struck me.. it's linear, i need log :(. So i connected up to a channel of my set-up anyway and gave it a proper test, it sounds good. But it's linear track is totally wrong, it doesn't have a fine enough adjustment for the 'quiet' end for night time listening and much too fine adjustment for the 'loud' end.
So i searched DIYaudio to see what i could find.. much the same here somebody suggested using a log lookup table, but that won't give me a finer adjustment, i could use -10db to -60db attenuators at the input then a buffer then -1db to -9db attenuators then another buffer which is a very good solution but u need too stages of amplification which i dont want (i wanna keep this project descrete SE) . Today i have been tumbling ideas around in my head all day :spin:. The best i can come up with is.....
Coupling a small stepper motor to a 10 turn or 22 turn cermet track trimmer or pot then getting some uControl to give me steps of adjustment. So with a 22 turn trimmer i should get over 1000 diffrent settings. Or i could use a 10 turn log pot, people say with dual log | | | |
