kevinkr said:
Parts come in today. Actually, they should be waiting when I get home... Then I'll just build up a simplified version the circuit on a breadboard, check how noisy the power supply design is, and tweak as needed. Hopefully the first run of five PCBs will be ordered by the middle of next week.
After receiving those and being sure that they work as expected I'll figure out what I'm going to do group buy wise.
Right now I'm trying to sort out an issue with having mute state saved between power cycles. I'm starting to think that there is a bug in mikroBasic 6 or the new PIC FLASH 2 programmer. Things just aren't working right immediately after programming, almost as if EEPROM is being written even when it isn't supposed to be.
I've tweaked the board design a little bit more and I'm quite happy with it. I added SMT jumpers on the bottom of the board for those who don't want to use wire bridges or actual jumpers for board config. I also added a fifth mounting hole, right next to the heatsink, so those of us who are sliding this into a Hammond 1455T enclosure will be able to fix the board to the bottom of the enclosure and keep it from sliding around. See here if you are curious.
So, it's all just in process... I hope to have more info later tonight or tomorrow.
-Steve
An externally hosted image should be here but it was not working when we last tested it.
Just an update for everyone:
Things are coming along nicely. Everything is working great. I've yet to test the power supply, but tonight's poking with the breadboard and tweaking the board layout is making things look nearly done.
That setup up above there (power LED and indicator LED on 150Ω resistors) was drawing 66mA of 5VDC, with the LEDs accounting for around 34mA of that. My meter registered a minimum of ~8mA when switching between relays was happeneing (likely in the 100ms when no relays are closed and no LEDs lit). So, this doesn't take much power at all.
I tossed a PDF of the schematic up, if anyone wants to see it: nuxx_audio_input_switch_draft_1.pdf
Of course, it's not done, but it's getting there...
I'll probably finalize the board design tomorrow, confirm the power supply works as I like, and order the first run of five boards (which are all accounted for) on Thursday.
Goodnight...
-Steve
kevinkr said:
I might not be understanding what you're saying, but do you mean to take the +/- from a source and tie it to its own ground/common, or tie it to chassis ground? I can't imagine tying it to its own ground/common would be a good idea... You'd be then basically tying everything coming out of the various source components all together.
If I'm not understanding your meaning, please correct me. It's the end of the work day here and I'm not thinking quite straight.
Looking at the board, I don't think there is room for something of this sort on the board. As it is, things are pretty tight...
-Steve
c0nsumer said:
I might not be understanding what you're saying, but do you mean to take the +/- from a source and tie it to its own ground/common, or tie it to chassis ground? I can't imagine tying it to its own ground/common would be a good idea... You'd be then basically tying everything coming out of the various source components all together.
If I'm not understanding your meaning, please correct me. It's the end of the work day here and I'm not thinking quite straight.
Looking at the board, I don't think there is room for something of this sort on the board. As it is, things are pretty tight...
-Steve
I think what he wants is to have no signal on the input sides of the relays that are off, so that the PCB traces and any coupling across the relay contacts is moot.
I would not recommend doing this if you don't know the design of the output stage on the source component, or without a high impedance input. I fried one channel of a DAC yesterday using it as a test source while prototyping (very similar project actually, signal-detecting source selector). Connecting the DAC output to a short through the microcontroller's protection diodes (oops, forgot that the output would go negative...) killed it. Probably not worth the risk - you'd need series resistors to protect against this, and of course that would kill the output impedance of the source.
At least, I think that was the failure mode. Either way, it's definitely a bad idea, you'd be shorting the AC signal directly to ground, and not all output stages would enjoy that.
error401 said:
Most properly designed sources should be short circuit proof and commercial practice is to use a small series resistor on the output to the rca jack. All current audio op-amps have short circuit protection - transistorized discrete class A se output circuits commonly found in old tuners and the like are inherently short proof. Balanced sources usually have 300 ohm resistors in series with the outputs.. Note again that I indicated this was to be something that could be configured as an option by the user/builder. This was actually fairly common practice in consumer audio electronics, and imo is still likely to be encountered from time to time.
The shorting configuration would normally be for use with single ended inputs to the input's ground in the event that you are switching those. (Good for ground loop avoidance.) It would be something that could be implemented or not as desired.
I am assuming that the dac you are referring to is a voltage output dac (chip) and the internal op-amp in most of these devices is not protected against shorts.
c0nsumer said:
I might not be understanding what you're saying, but do you mean to take the +/- from a source and tie it to its own ground/common, or tie it to chassis ground? I can't imagine tying it to its own ground/common would be a good idea... You'd be then basically tying everything coming out of the various source components all together.
If I'm not understanding your meaning, please correct me. It's the end of the work day here and I'm not thinking quite straight.
Looking at the board, I don't think there is room for something of this sort on the board. As it is, things are pretty tight...
-Steve
Hi c0nsumer, kevinkr, would this link help in regards too...
"configure the inputs such that unselected inputs can be shorted to ground at the board user's discretion to reduce cross-talk from unselected, but active sources."
The schematic came from Morgan Jones from a discussion on
shunt and series switching using relays eliminates crosstalk.
http://www.diyaudio.com/forums/showthread.php?postid=786619#post786619
Stan
sklimek said:
If I'm reading that right, that'd double the number of relays needed. That breaks one of my initial design requirements of fitting this inside one of the preexisting Hammond cases along with my Millet Hybrid Maxed. :\
So, I don't think I'm going to be able to do that... At least not that way...
Also, I am intending to also use this as a SPDIF switch, so I'd have to have that (as mentioned before) as a configurable option... That'd add even more parts to the board. So, I think I'm going to stick with it as it is...
For what it's worth, I'm just finishing out changes to the power supply and routing things on the board. On the scope I see a nice, solid 5V with no glitches when switching inputs. Scope and meter show around 3mV AC ripple, which may very well be picked up by the long leads.
I've also got the regulator in the breadboard, wrapped in masking tape to hold a thermocouple to the back side of it. Steady state (~1 hour powered on) with Input 1 selected and a power LED (just simple red ones with 150 ohm resistors) I'm seeing the reg as being at 128.7F. The assembly is powered via bridge rectifier, which is receiving 13.4VAC. That temp is well within operating specs, so, I think with the regulator clamped to the heatsink everything should be just fine.
Hi C0nsumer,
Mine was a suggestion only, the current implementation should be as good or better than what I currently have. I think the additional relays would be a PITA to fit into the available space.
I've had problems switching spdif in the past, you need to make sure that you use low capacitance relays to minimize cross-talk. In this case a T style switching arrangement might be advised, but try what you've designed for the moment - this will in any case be a very useful audio switcher and with some care will probably work well for spdif as well.
Mine was a suggestion only, the current implementation should be as good or better than what I currently have. I think the additional relays would be a PITA to fit into the available space.
I've had problems switching spdif in the past, you need to make sure that you use low capacitance relays to minimize cross-talk. In this case a T style switching arrangement might be advised, but try what you've designed for the moment - this will in any case be a very useful audio switcher and with some care will probably work well for spdif as well.
kevinkr said:
I understand, thanks for at least mentioning it. It got me thinking about some other things too.
I really appreciate any suggestions, because they get me thinking about possible ways to do things differently.
For what it's worth, I'm just finishing the board design now. My experiences in testing the power supply last night showed that I needed a much larger smoothing cap than expected, especially if I want a relatively stable 5V output available when the board is being fed 6.2VAC. I've managed to squeeze in a 13mm footprint (5mm LS) and I'm just sorting out the remaining details.
I'm still hoping to order the first run of boards this evening. I'll post an image of the board after work or something.
-Steve
An externally hosted image should be here but it was not working when we last tested it.
Well, the first five PCBs were ordered from PCBEX.COM a few minutes ago. The biggest changes since yesterday were moving the smoothing capacitor before the voltage regulator to 1000uF (13mm diameter, 5mm lead spacing footprint) which required a good bit of shuffling of parts in the power supply. (Some configurations, particularly those not needing to use the 5VDC output, can use a much smaller cap here. I'm recommending this value as a best-fit.) As part of this I also ended up moving all of the connectors to the front edge of the board, which I think is a rather good thing.
The boards are five-day turn, so I figure they'll ship sometime around the end of next week. As PCBEX.COM batches their boards out to China I expect it'll be another week or so before I have the boards.
Here's how the part prices from Mouser breaks down:
All parts, except PCB, heatsink, relays, and audio terminals: US$10.46
5x 4PDT Omron relays and 5 place Phoenix terminals: US$57.10
5x DPDT Omron relays and 3 place Phoenix terminals: US$25.45
1" Tall Heatsink: US$1.14
1.5" Tall Heatsink: US$1.06
Of course, you can cut these prices down if you do things such as not fitting the boards to switch all five inputs, leave out some of the onboard power supply parts, etc.
The PCBs cost about US$16/each after shipping from PCBEX.COM. If I were purchasing for a group buy or something I imagine the price would be a good bit lower, although that obviously depends on the quantity.
So, I guess that's about it for now... There is more info over at the main nuxx Audio Input Switch (as the device is now known) page.
One final question for everyone: Are any of you interested in purchasing some of these PCBs in a group buy?
I figure that I could probably handle doing a group buy for some folks. I'd provide a PCB and programmed PIC, and you'd just have to order the rest of the parts. Everything in there is very standard, so even if you can't order from Mouser you shouldn't have a problem sourcing the same (or compatible parts. I'm not in the position to offer kits or assembled boards.
Off the top of my head I'd estimate the total cost at being around $20, to cover the blank PCB, PIC, packing stuffs, PayPal fees, etc.
Anyway, that's it for now...
-Steve
Where can I find the schematic ? Looking at the PCB I think you will have crosstalk as signal carrying traces are very close to eachother. A ground track in between would be better. Relays are also close to eachother, some more space would be better to reduce crosstalk. This could be realised by moving the outputs to the right side of the PCB. Just some suggestions.
Also grounding the unused inputs would be a nice option. When using the PCB unbalanced this can be done by grounding the middle row of contacts with a straight wire to signal ground.
Apart from that a nice looking PCB. I would buy one as I already have the relays. If you know the Opus thread you won' t want me to do beta testing !
Also grounding the unused inputs would be a nice option. When using the PCB unbalanced this can be done by grounding the middle row of contacts with a straight wire to signal ground.
Apart from that a nice looking PCB. I would buy one as I already have the relays. If you know the Opus thread you won' t want me to do beta testing !
jean-paul said:
Sorry, here is the schematic. I haven't updated the page with it yet.
One thing, though, won't crosstalk only be a problem if there is actually a signal going into the other relays? Personally, I normally switch off (otherwise stop output) from the devices I'm not listening to, but I guess in the case of things like a TiVo that might not be possible...
The physical size is actually the result of a design constraint so that it can fit in a Hammond 1455T enclosure I already have. It's also maxed out the EAGLE non-profit license I have, as 160mm is the board size limit on the X axis. The only way I could better space the relays is to drop down to switching only four inputs.
Even though the boards are ordered, I was giving a quick look as to how I could ground the inputs... I can see a reasonable way to do it, but I can't currently think of a good way to make it builder selectable without the cutting of traces. So... hrm...
As far as selling the boards, thanks for the interest. I'm just going to hold off on the group buy stuff until I'm sure that the boards will work, but I am really glad to see that there's interest. Once everything is working I'll try and get the board info to make the normal DIY electronics rounds to get as much interest as possible.
-Steve
Dropping one relay will give better opportunities for making the design of better quality. Or making the PCB bigger but with the need for buying an Eagle license. I would also incorporate resistors from inputs to ground to prevent plops ( reference to ground) and series resistors in the output to the volume control. As it is now the contacts are floating which will definitely result in clicks and plops while switching.
A low pass filter at the inputs consisting a a resistor and a small cap to avoid RF coming in would be the icing on the cake. Or at the output to keeps things simple. Let's say around 100 kHz to avoid any filtering in the audioband when sources with varying impedance are switched.
If you don't like the low pass filter please realise that if you plan to ground the unused inputs a series resistor is required of let's say 47 Ohm. Shorting is not the best method with DC coupled sources The low pass filter would have a double function so to speak.
A low pass filter at the inputs consisting a a resistor and a small cap to avoid RF coming in would be the icing on the cake. Or at the output to keeps things simple. Let's say around 100 kHz to avoid any filtering in the audioband when sources with varying impedance are switched.
If you don't like the low pass filter please realise that if you plan to ground the unused inputs a series resistor is required of let's say 47 Ohm. Shorting is not the best method with DC coupled sources
The low pass filter would have a double function so to speak.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.