·· It looks like you've found someone as passionate (Or more) about their audio as I am!· LOL· In any event, I have to admit that my current project is much more low-key than anything I used to build.· Since it's now just for my personal use I have no need currently for lots of power, just quality.· But there may come a day when I require a little more boost, and we'll see what the current market holds for me then.· Of course you know it will be Stamp-Controlled...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
Yeah, I admit, I can get a little bit zealous about audio, mostly as the result of a lawsuit by B*se, suing me for publishing frequency response graphs of the original AM-5. I've been something of a crusader for debunking and truth ever since. I've towned it down a lot, but there is still so much misinformation out there that I'll never be satisfied. As for music, the industry almost make sme cry over what they've done to the sound in order to get something louder on the radio, or the lack of diversity in programming.
But enough about that, back to audio.
Payson-
The Tripath stuff is something I've been following for a number of years, and it has promise, but I haven't managed to toy with it enough to make myself happy with the result. They have an interesting way of varying the PWM frequewncy depending on the output to supposedly control dissipation as well as mniimize filtering artifacts. It mostly works, and I'm in the process of building another pair of modules that should provide a much more rewarding result. The kist are a product of a nice Swede named Jan and all the info is available at his website (http://www.41hz.com). The AMP2 is the kit I'm working with. Jan and I have been bouncing back and forth about the possibility of a very large amplifier based on similar topologies...8kW or more, for professional applications. It's a slow road (building an 8+ kW power factor corrected 3 phase internationally compliant power supply is a real task), but eventually we might just get it done.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· I was wondering if you or anyone else knew of any decent audio switching ICs that work with dual-voltage power supplies.· I've been doing a lot of research lately and coming up with only single-supply units or obsolete chips, or, in a few cases, surface mount only devices.
·· I'm trying to find a good quality audio switching chip for my next audio project that can switch inputs based on I2C, SPI or parallel input control.· Any ideas?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
The only ones I've found and used are the muxs from TI (nee Burr-Brown). I experimented way back when with some oddball topologies for using small signal FETs back-to-back for switching at my desired max rail voltages (+/-24VDC) but I could never match them well enough to make things as clean as I hoped, and control was rough. The MPC507 was the best one I could come up with. I dug up the Intersil DG407 a few weeks ago, and some samples are supposedly winging their way to me now, but it looks more like second source of the MPC507 (or maybe the other way around?) than anything else. They both use a 1-of-8 or 1-of-16 (depending on if it's dual 1:8 or single 1:16 mux) parallel control with an enable pin, and both have a +/-15Vinput range (with +/-15-18V rails IIRC) . Maxim has a couple of SPI-to-parallel output expanders I've used that would work great for controlling a big array of MUXs.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· You're right, that one is a replacement for the 507...I do appreciate the information though.· In a recent Nuts & Volts article I saw what I thought was a new replacement chip which would be great for the job, but it is I think 5V single-supply.· I don't even know how they pass that stuff of as audio applications...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
I kinda figured they were mutual sources after I stared at the datasheets side by side. Silly me for not noticing that earlier. I've often thought about using some small relays for the task, but the kind I want to use are much too expensive to replace the BB/Intersil parts (vacuum chamber, gold contact, zero starting current, shielded DC coil). When I use them as balanced switches (2x[noparse][[/noparse]8x1]), they seem to match quite well. Balancing across substrates (2x[noparse][[/noparse]16x1]) makes it harder, as thermal gradients and manufacturing tolerances crop up to clobber my THD numbers. My personal best is 0.006%THD+N with a carefully designed board and triple filtered rails, tho it's too much of a space/heat hog to use in my 4th gen preamp/integrated amp. Unfortunate really.
What's the project you envision using the 507/407s for?
An amazing amount of things are being shoved into the 0-5V analog space, and like you, it drives me slightly nuts. I recently had a trade pub slide across my desk (on it's way to the circular file) and on one of the back pages was an ad for a true differential ADC with incredible noise and SFDR specs with some kind of wildly low price. Of course, it was a 0-5V piece, whose application I can only guess at. A lot of it is being driven by the transition to lower voltages for logic and the swelling of the battery and switchmode markets. If you've already got a stiff 5V supply floating around for other things, why not just add caps and call it good for audio? 'A zero-and-span at the end of the chain isn't bad' is something I've heard from customers on more than one occasion, and it gives me fits. Maybe you can do it in a walkman with a little 100mW BTL headphone amp, since there's only a couple of AAs powering the beast, but seriously, in high end gear? Ugh.
In short, I feel your pain. High quality bipolar analog is being pushed into a very small niche, populated only by the highest end stereo companies and a subset of the pro audio market. It's sad really.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· I shouldn't complain...My original amplifier design from 1992 used the LM1037 which was single-supply, and I got used to using it.· But I prefer to have the dual-supply support since my pre-amps and main amp are dual-supply.· It's really just a matter of perspective/asthetics from the stand-point of consistency with the structure of the signal going through the amplifier.
·· In any event, the LM1037 was discontinued, which cause my move to other methods of switcing the input, of which I have even tried/used relays (PCB-Mount).· I guess I was looking for a nice solution to switch between 4 and 8 stereo inputs into a new amp I am building.· Like all my other units, this one will feature a separate Audio Controller section/module, which will contain all the logic and control (Including switching audio) and a separate power amp section, which will be nothing more than a cabinet with the power supply for the amp, a power meter on the front, and the connections in the back for the incoming audio from the controller.· There will also be a fan control system.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
Post Edited (Chris Savage (Parallax)) : 9/19/2005 11:20:47 PM GMT
Let me recommend the 407/507 highly then. A single chip will do your stereo 8x1 switching in one box, and they're stupid-easy to use.
I'm actually moving in the opposite direction with my system. I ahve a verynice seperates system for the living room, but I want sound in the bedroom to be a little more compact, so I'm building an integrated amp, with my pre sections (with some rework) and a Tripath based amp that's ~ 250Wx2 at 8R. I'm trying to come up with a good way to modularize my preamp subsystems so I can eventually build it into a whole-house audio system, with multi-source selection and LCD keypads, but that's a little ways off I think. I'm still waiting to see where Xantech goes with their big 8 zone mainframe (MRC88). I absolutely love their hardware for the LCD touchpanels too, but the programming system leaves a lot to be desired.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· 250W X 2?· For a bedroom?· Wow!· Anyway, I took your advice and ordered a few samples of that chip.· I have no real information that lists it's audio bandwidth characteristics.· So I hope it's pretty good.· I will be using it with the LM3886TF.· On the pre-amps, I forgot what I was using, but I will go back and re-design the whole front end once I've decided whether I want electronic control of the volume (I'm thinking not at this point).
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
I'm all about headroom and dynamic range. The speakers for the bedroom are also horribly inefficient at 87dB 1w/1m, and having the current dump capabality of 250W makes the bass nice and tight and the lower mids crisp, to use some audiophoole jargon.
The 3886 is a great little part in all it's incarnations. I have a pile of them, both in projects and still in the rails. Let me know if you run out. National has gotten pretty stingy with sampling them, especially the -TF parts, recently.
My $0.02:
Digital control of the volume is pretty easy, especially if you use a rotary encoder to do the adjustments like is shown in the SX/B examples. Scale 0-100% as 0-192 and dump it directly into a Crystal CS3310 (or a TI PGA2310) as a hex level value, shazam, you're done. If you build it like I did so that the volume change is handled (more or less) in the main loop and reacts only when the setpoint is different from the current point, you can control it from an encoder (ISR, just bump the setpoint as the knob turns) or from an IR remote just as easily, and seamlessly I might add. Once you've got a remote, it's really hard to go back. Really hard.
The current state of my preamp control project in c-like pseudocode:
ISR_read_encoder()
{
read encoder
}
ISR_oops()
{
fault management
}
main()
{
read IR remote
read front panel UI
if (mute changed) update muting
if (volume changed) update volume
if (sourse changed) update source
if (anything the UI shows changed) update display
feed watchdog
}
UI_settings_handler()
{}
/* big ugly subroutine that allows me to do things like change source names, change tweak values for source volume, change timeouts, brightness levels, remote response speed, etc */
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· Thanks for the advice on the Crystal chip...I haven't forgotten about it.· I have the datasheet right there with the 507 in my Audio Datasheets folder.· The thing holding me back right now is my layout for the LM3886TF.· I attempted a layout in ExpressPCB without my original to use as a guideline.· It just doesn't look right.· I was originally hoping to make a small compact board that just holds the IC and support components.· Two of these would mount to the heatsink and their connections would go from there to the P/S and input jacks.· But that's not what I ended up doing.· I don't like what I have, and when I get more motivation I will try again.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
Regarding the LM3886, there's an Eagle library for it, and I can dig up one of my old ones and send it to you if you want. ExpressPCB is painful for me to use these days. I don't have any PCBs kicking around, but there is a PDF on my website that has info on a modular board I designed a while back, that might work out for you. The output inductor isn't onboard, but it should be at the output jacks anyway.It's also ugly, and requires hand-photoetching due to my completely unsophisticated way of building them back then. I'll see if I can bang out a better version this week in the off-hours. Is there anything you want to have on the board that isn't in the basic datasheet schematic? I prefer SMT things these days, so I'll probably lay it out with 0805s and 1206s for ease of assembly. I'll also see about making it modular so that it could be built in blocks of 4, and split as required.
Thoughts?
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· I can send you the ExpressPCB layout if you want to see what I was doing.· Mine did account for the inductor, since the ones I use are kind of made for PCB...In any event, my circuit varies slightly from the one on the datasheet for typical audio amplifier, but I don't recall off-hand the differences.· Again, I can check later.· As for surface mount, I haven't done any yet.· No motivation for that.· If you build a nice LM3886TF PCB, I'd be interested in checking it out, however I don't have Eagle yet, so...
·· I can tell you that I do know I want my controller to have a nice LCD, LED indicators for selection, etc.· However, I will most likely use a BS2px24 in it.· Possibly a BS2p40.· Both are fast enough to read a digital encoder and run the rest of the show as well.· The encoders I use have a push-button in them.· I have some more ideas I will get on paper this weekend.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
Fair enough. I'll PM you my real contact info for that ExpressPCB stuff. Let me know what else you'd like on the board. The PA-100 schematic in NS AN-1192 is one that I've been toying with for a while. It allows full single-chip output into 3-4 ohms with 38V rails, which is more flexible than the single chip solution. Something to consider.
I used a 64 step Grayhill encoder in my project and my BS2P was easily overloaded with a 2" knob on it. I have some chips that convert the quadrature signal from the encoder into up and down pulses with an interrupt out pin, which is perfect for what I want to do, as the encoder read interrupt will only be a line or two of SX/B, compiling to maybe 8 or 10 assembly instructions, making re-entrant interrupt concerns pretty much moot. I have a couple of SPI based serial expanders that are designed for button inputs, and also offer an interrupt out pin, again, perfect, since I can use small routines and pound the rest into if() updates in the main loop. Fun stuff.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· 64 Steps may be a problem...The ones I have are 16 steps, and seem to work great in all the other BS2p applications I found for them.· If you really try to spin the knob extremely fast the BS2p will miss a few clicks, but I never do that anyway.
·· I suppose doing this on an SX could be a consideration as well.· I have always used BASIC Stamps since 2000 in all of these projects, but we'll see what happens.· Still in the design phase on this one.··BTW, I sent you that ExpressPCB schematic to check out.· No schematic yet though.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
I'll see if I can cook up a layout this weekend for you. I'll have to dig out my old Thinkpad with ExpressPCB on it. My Mac doesn't play nice with .pcb files
The 64 step is mainly so I can traverse the 0-256 level range for the CS3310s quickly (my knob is a chunk of machined aluminium that looooves to spin). I've also got 16, 32, and 128 step versions, but the 64 was the one that 'felt' the best (steps vs knob diameter), until I added all the handlers to my code to make it work. The BS2 couldn't get it all done in time. I've been a Parallax guy since Chip posted to the mailing list and the BS-1 was brand-spanking-new, but the speed and flexability of the SXs is an undeniable siren song for me. One simple interrupt with the external chip and I'll be set. It should be fun.
FWIW, the encoder interface chips are from LSI products.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· One thing I will need to do is check my current version of ExpressPCB.· I have been hearing a lot of talk on here about a new version with added features, and I am not sure how new mine is.· I will be updating it this weekend as well.· Thanks for your help and input.· Of all the Amp Boards I have built, I want this one to outshine them all before I commit to it.·· I hope to have it for a very long time.
·· I also need to find a nice case.· I have always used cheap/inexpensive cases for prototyping, but I lose the ability to have nice built-in heat sinks as I see your cases have.· I think I better find a good deal on one soon.· Especially since I have nothing here large enough to prototype this one anyway.· At least if I get a new case I can start on milling the front and read for the parts that ARE set, such as the meters and terminals.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
I think you may have overlooked something. The cases that I showed in my 3886 project page were Radio Shack specials. The heatsinks were slavaged from a piece of surplus equipment and I did come cutting and grinding to attach them to the existing enclosure. Even in my commercial designs, it's rare that I get to use a case with anything close to integrated heatsinking for a product. 99.9% of the time, the sinks are either inside the box with holes punched, or they're mounted outside and the case is modified as my amp cases were.
If you look carefully at this image, you can see the hex cap screws I used to secure the sink to the back of the enclosure. I cut a slot to allow the big mounting tap on the sink to come through the case.
I'm unaware of any ExpressPCB changes. Then again, I haven't used it in a few years, so maybe I should update my copy
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· You're right, in fact one of the enclosures you're using matches one I have from Radio Shack, however, the one with the Amp channels mounted on either side looks like a more professional unit that I have seen while searching LM3886 amplifiers on Google.· I am thinking of using something like this for my amp.
·· Where is that Crystal chip available from?· I tried Digikey, Mouser and Jameco, and there was no match on the part#.· Granted, I didn't try everything, but where do you get them from?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
Well there's the LM12 amp, which was also assembled from heatsink stock, but was turned into a rack-like enclosure with the aid of some aluminum L-bar and a steel channel. There's also my collection of Haflers, which use custom-formed cases and some semi-custom (order 3000ft to get any at all) Aavid extrusions. You might check out the folks at Par Metals too. They're a favorite of the DIY-pro audio commmunity. The 20 series looks like it might work for you. www.par-metal.com
The Crystal chip is made by Cirrus Logic now (they bought Crystal a while back) and is much more easily purchased as a TI PGA2311. The PGA2320 might be a better option, given it's better matching, lower THD, and higher voltage rails (+/-15VDC, vs the 2311s and CS3310s +/-5V)
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
Post Edited (Dave Paton) : 9/25/2005 3:19:49 AM GMT
·· Thanks for the information.· I think I will order the enclosures next.· As for the replacement chip, I noticed right away that it's SOIC.· It will be interesting to see me get into surface mount components finally.· I think to start I will try the CS3310, but I am ordering the other one for later use.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
I'd get a sample of the PGA2311. It's literally a drop-in replacement, and much more readily available than the Crystal/Cirrus part, mainly because TI samples them, which allows you to avoid dealing with Newark (blech). I've effectively left Cirrus behind for my newer designs, because of the easier availability of the TI cross and the friendliness of the TI dealers and saple reps.
My $0.02
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· Your advice is taken...Tonight when I get home I will sample that part and start on a design for the PCB.· I may want to find at least one of the old part in DIP format to experiment with first though so I can more easily tweak the circuit prior to committing to PCB.
·· Hey, speaking of the LM12...Did you ever find out why yours developed the hum?· I wondered after looking at the schematic if it was related to a faulty part on your bridge on the output.· That's the only place I could think of from just looking at the schematic.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
The hum was actually a result of not having top and bottom panels, combined with a really crappy crossover I was using. In hindsight, I should have built it a lot better. That mess of wire in the middle of the amp is frightening to me these days, knowing what I now know about noise and induction and things of that sort. Ugly.
There are some versions of the PGA parts in DIP IIRC. The CS3310 only ever came in a SOIC.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
No worries. I've missed more obvious things to much greater detriment.
The Phase Linear 700 II, Crown MT1200, Crown D75, Crown K2, and Lab.Gruppen fp6400 all make up for the loss of the LM12 amp. I may resurrect it, sorta, in a pair of powered speakers I'm designing. Then again, I might just use one of the (many) amps that's kicking around and collecting dust. We'll see.
Yes, my house does have twist-loks for the stereo. I know I'm crazy.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
Okay, so by this weekend I will have a cabinet layout draft done.· I now know, based on the parts I will be using exactly what buttons, displays, LEDs, etc. I will need, and I am just working out where I want them.· I will try to post a graphic by this weekend on it, since as soon as I get a case, I will start milling and working on board/part placement.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Chris Savage Parallax Tech Support csavage@parallax.com
I'm looking forward to seeing what you do with your project. Mine is progressing slowly, but a visit from the friendly Noritake rep left me with some amazing samples of some VFDs and I'm looking forward to working on it more and more.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
Comments
·· It looks like you've found someone as passionate (Or more) about their audio as I am!· LOL· In any event, I have to admit that my current project is much more low-key than anything I used to build.· Since it's now just for my personal use I have no need currently for lots of power, just quality.· But there may come a day when I require a little more boost, and we'll see what the current market holds for me then.· Of course you know it will be Stamp-Controlled...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
Yeah, I admit, I can get a little bit zealous about audio, mostly as the result of a lawsuit by B*se, suing me for publishing frequency response graphs of the original AM-5. I've been something of a crusader for debunking and truth ever since. I've towned it down a lot, but there is still so much misinformation out there that I'll never be satisfied. As for music, the industry almost make sme cry over what they've done to the sound in order to get something louder on the radio, or the lack of diversity in programming.
But enough about that, back to audio.
Payson-
The Tripath stuff is something I've been following for a number of years, and it has promise, but I haven't managed to toy with it enough to make myself happy with the result. They have an interesting way of varying the PWM frequewncy depending on the output to supposedly control dissipation as well as mniimize filtering artifacts. It mostly works, and I'm in the process of building another pair of modules that should provide a much more rewarding result. The kist are a product of a nice Swede named Jan and all the info is available at his website (http://www.41hz.com). The AMP2 is the kit I'm working with. Jan and I have been bouncing back and forth about the possibility of a very large amplifier based on similar topologies...8kW or more, for professional applications. It's a slow road (building an 8+ kW power factor corrected 3 phase internationally compliant power supply is a real task), but eventually we might just get it done.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· I was wondering if you or anyone else knew of any decent audio switching ICs that work with dual-voltage power supplies.· I've been doing a lot of research lately and coming up with only single-supply units or obsolete chips, or, in a few cases, surface mount only devices.
·· I'm trying to find a good quality audio switching chip for my next audio project that can switch inputs based on I2C, SPI or parallel input control.· Any ideas?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
The only ones I've found and used are the muxs from TI (nee Burr-Brown). I experimented way back when with some oddball topologies for using small signal FETs back-to-back for switching at my desired max rail voltages (+/-24VDC) but I could never match them well enough to make things as clean as I hoped, and control was rough. The MPC507 was the best one I could come up with. I dug up the Intersil DG407 a few weeks ago, and some samples are supposedly winging their way to me now, but it looks more like second source of the MPC507 (or maybe the other way around?) than anything else. They both use a 1-of-8 or 1-of-16 (depending on if it's dual 1:8 or single 1:16 mux) parallel control with an enable pin, and both have a +/-15Vinput range (with +/-15-18V rails IIRC) . Maxim has a couple of SPI-to-parallel output expanders I've used that would work great for controlling a big array of MUXs.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· You're right, that one is a replacement for the 507...I do appreciate the information though.· In a recent Nuts & Volts article I saw what I thought was a new replacement chip which would be great for the job, but it is I think 5V single-supply.· I don't even know how they pass that stuff of as audio applications...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
I kinda figured they were mutual sources after I stared at the datasheets side by side. Silly me for not noticing that earlier. I've often thought about using some small relays for the task, but the kind I want to use are much too expensive to replace the BB/Intersil parts (vacuum chamber, gold contact, zero starting current, shielded DC coil). When I use them as balanced switches (2x[noparse][[/noparse]8x1]), they seem to match quite well. Balancing across substrates (2x[noparse][[/noparse]16x1]) makes it harder, as thermal gradients and manufacturing tolerances crop up to clobber my THD numbers. My personal best is 0.006%THD+N with a carefully designed board and triple filtered rails, tho it's too much of a space/heat hog to use in my 4th gen preamp/integrated amp. Unfortunate really.
What's the project you envision using the 507/407s for?
An amazing amount of things are being shoved into the 0-5V analog space, and like you, it drives me slightly nuts. I recently had a trade pub slide across my desk (on it's way to the circular file) and on one of the back pages was an ad for a true differential ADC with incredible noise and SFDR specs with some kind of wildly low price. Of course, it was a 0-5V piece, whose application I can only guess at. A lot of it is being driven by the transition to lower voltages for logic and the swelling of the battery and switchmode markets. If you've already got a stiff 5V supply floating around for other things, why not just add caps and call it good for audio? 'A zero-and-span at the end of the chain isn't bad' is something I've heard from customers on more than one occasion, and it gives me fits. Maybe you can do it in a walkman with a little 100mW BTL headphone amp, since there's only a couple of AAs powering the beast, but seriously, in high end gear? Ugh.
In short, I feel your pain. High quality bipolar analog is being pushed into a very small niche, populated only by the highest end stereo companies and a subset of the pro audio market. It's sad really.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· I shouldn't complain...My original amplifier design from 1992 used the LM1037 which was single-supply, and I got used to using it.· But I prefer to have the dual-supply support since my pre-amps and main amp are dual-supply.· It's really just a matter of perspective/asthetics from the stand-point of consistency with the structure of the signal going through the amplifier.
·· In any event, the LM1037 was discontinued, which cause my move to other methods of switcing the input, of which I have even tried/used relays (PCB-Mount).· I guess I was looking for a nice solution to switch between 4 and 8 stereo inputs into a new amp I am building.· Like all my other units, this one will feature a separate Audio Controller section/module, which will contain all the logic and control (Including switching audio) and a separate power amp section, which will be nothing more than a cabinet with the power supply for the amp, a power meter on the front, and the connections in the back for the incoming audio from the controller.· There will also be a fan control system.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
Post Edited (Chris Savage (Parallax)) : 9/19/2005 11:20:47 PM GMT
Let me recommend the 407/507 highly then. A single chip will do your stereo 8x1 switching in one box, and they're stupid-easy to use.
I'm actually moving in the opposite direction with my system. I ahve a verynice seperates system for the living room, but I want sound in the bedroom to be a little more compact, so I'm building an integrated amp, with my pre sections (with some rework) and a Tripath based amp that's ~ 250Wx2 at 8R. I'm trying to come up with a good way to modularize my preamp subsystems so I can eventually build it into a whole-house audio system, with multi-source selection and LCD keypads, but that's a little ways off I think. I'm still waiting to see where Xantech goes with their big 8 zone mainframe (MRC88). I absolutely love their hardware for the LCD touchpanels too, but the programming system leaves a lot to be desired.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· 250W X 2?· For a bedroom?· Wow!· Anyway, I took your advice and ordered a few samples of that chip.· I have no real information that lists it's audio bandwidth characteristics.· So I hope it's pretty good.· I will be using it with the LM3886TF.· On the pre-amps, I forgot what I was using, but I will go back and re-design the whole front end once I've decided whether I want electronic control of the volume (I'm thinking not at this point).
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
I'm all about headroom and dynamic range. The speakers for the bedroom are also horribly inefficient at 87dB 1w/1m, and having the current dump capabality of 250W makes the bass nice and tight and the lower mids crisp, to use some audiophoole jargon.
The 3886 is a great little part in all it's incarnations. I have a pile of them, both in projects and still in the rails. Let me know if you run out. National has gotten pretty stingy with sampling them, especially the -TF parts, recently.
My $0.02:
Digital control of the volume is pretty easy, especially if you use a rotary encoder to do the adjustments like is shown in the SX/B examples. Scale 0-100% as 0-192 and dump it directly into a Crystal CS3310 (or a TI PGA2310) as a hex level value, shazam, you're done. If you build it like I did so that the volume change is handled (more or less) in the main loop and reacts only when the setpoint is different from the current point, you can control it from an encoder (ISR, just bump the setpoint as the knob turns) or from an IR remote just as easily, and seamlessly I might add. Once you've got a remote, it's really hard to go back. Really hard.
The current state of my preamp control project in c-like pseudocode:
ISR_read_encoder()
{
read encoder
}
ISR_oops()
{
fault management
}
main()
{
read IR remote
read front panel UI
if (mute changed) update muting
if (volume changed) update volume
if (sourse changed) update source
if (anything the UI shows changed) update display
feed watchdog
}
read_buttons()
{}
read_IR()
{}
update_mute()
{}
update_volume()
{}
update_source()
{}
update_display()
{}
UI_settings_handler()
{}
/* big ugly subroutine that allows me to do things like change source names, change tweak values for source volume, change timeouts, brightness levels, remote response speed, etc */
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· Thanks for the advice on the Crystal chip...I haven't forgotten about it.· I have the datasheet right there with the 507 in my Audio Datasheets folder.· The thing holding me back right now is my layout for the LM3886TF.· I attempted a layout in ExpressPCB without my original to use as a guideline.· It just doesn't look right.· I was originally hoping to make a small compact board that just holds the IC and support components.· Two of these would mount to the heatsink and their connections would go from there to the P/S and input jacks.· But that's not what I ended up doing.· I don't like what I have, and when I get more motivation I will try again.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
Thoughts?
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· I can send you the ExpressPCB layout if you want to see what I was doing.· Mine did account for the inductor, since the ones I use are kind of made for PCB...In any event, my circuit varies slightly from the one on the datasheet for typical audio amplifier, but I don't recall off-hand the differences.· Again, I can check later.· As for surface mount, I haven't done any yet.· No motivation for that.· If you build a nice LM3886TF PCB, I'd be interested in checking it out, however I don't have Eagle yet, so...
·· I can tell you that I do know I want my controller to have a nice LCD, LED indicators for selection, etc.· However, I will most likely use a BS2px24 in it.· Possibly a BS2p40.· Both are fast enough to read a digital encoder and run the rest of the show as well.· The encoders I use have a push-button in them.· I have some more ideas I will get on paper this weekend.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
Fair enough. I'll PM you my real contact info for that ExpressPCB stuff. Let me know what else you'd like on the board. The PA-100 schematic in NS AN-1192 is one that I've been toying with for a while. It allows full single-chip output into 3-4 ohms with 38V rails, which is more flexible than the single chip solution. Something to consider.
I used a 64 step Grayhill encoder in my project and my BS2P was easily overloaded with a 2" knob on it. I have some chips that convert the quadrature signal from the encoder into up and down pulses with an interrupt out pin, which is perfect for what I want to do, as the encoder read interrupt will only be a line or two of SX/B, compiling to maybe 8 or 10 assembly instructions, making re-entrant interrupt concerns pretty much moot. I have a couple of SPI based serial expanders that are designed for button inputs, and also offer an interrupt out pin, again, perfect, since I can use small routines and pound the rest into if() updates in the main loop. Fun stuff.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· 64 Steps may be a problem...The ones I have are 16 steps, and seem to work great in all the other BS2p applications I found for them.· If you really try to spin the knob extremely fast the BS2p will miss a few clicks, but I never do that anyway.
·· I suppose doing this on an SX could be a consideration as well.· I have always used BASIC Stamps since 2000 in all of these projects, but we'll see what happens.· Still in the design phase on this one.··BTW, I sent you that ExpressPCB schematic to check out.· No schematic yet though.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
I'll see if I can cook up a layout this weekend for you. I'll have to dig out my old Thinkpad with ExpressPCB on it. My Mac doesn't play nice with .pcb files
The 64 step is mainly so I can traverse the 0-256 level range for the CS3310s quickly (my knob is a chunk of machined aluminium that looooves to spin). I've also got 16, 32, and 128 step versions, but the 64 was the one that 'felt' the best (steps vs knob diameter), until I added all the handlers to my code to make it work. The BS2 couldn't get it all done in time. I've been a Parallax guy since Chip posted to the mailing list and the BS-1 was brand-spanking-new, but the speed and flexability of the SXs is an undeniable siren song for me. One simple interrupt with the external chip and I'll be set. It should be fun.
FWIW, the encoder interface chips are from LSI products.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· One thing I will need to do is check my current version of ExpressPCB.· I have been hearing a lot of talk on here about a new version with added features, and I am not sure how new mine is.· I will be updating it this weekend as well.· Thanks for your help and input.· Of all the Amp Boards I have built, I want this one to outshine them all before I commit to it.·
·· I also need to find a nice case.· I have always used cheap/inexpensive cases for prototyping, but I lose the ability to have nice built-in heat sinks as I see your cases have.· I think I better find a good deal on one soon.· Especially since I have nothing here large enough to prototype this one anyway.· At least if I get a new case I can start on milling the front and read for the parts that ARE set, such as the meters and terminals.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
I think you may have overlooked something. The cases that I showed in my 3886 project page were Radio Shack specials. The heatsinks were slavaged from a piece of surplus equipment and I did come cutting and grinding to attach them to the existing enclosure. Even in my commercial designs, it's rare that I get to use a case with anything close to integrated heatsinking for a product. 99.9% of the time, the sinks are either inside the box with holes punched, or they're mounted outside and the case is modified as my amp cases were.
If you look carefully at this image, you can see the hex cap screws I used to secure the sink to the back of the enclosure. I cut a slot to allow the big mounting tap on the sink to come through the case.
I'm unaware of any ExpressPCB changes. Then again, I haven't used it in a few years, so maybe I should update my copy
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· You're right, in fact one of the enclosures you're using matches one I have from Radio Shack, however, the one with the Amp channels mounted on either side looks like a more professional unit that I have seen while searching LM3886 amplifiers on Google.· I am thinking of using something like this for my amp.
http://cgi.ebay.com/Tube-Amp-Chassis-Electronic-Enclosure-Ca_W0QQitemZ7547689713QQcategoryZ4660QQcmdZViewItem
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
·· Where is that Crystal chip available from?· I tried Digikey, Mouser and Jameco, and there was no match on the part#.· Granted, I didn't try everything, but where do you get them from?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
Well there's the LM12 amp, which was also assembled from heatsink stock, but was turned into a rack-like enclosure with the aid of some aluminum L-bar and a steel channel. There's also my collection of Haflers, which use custom-formed cases and some semi-custom (order 3000ft to get any at all) Aavid extrusions. You might check out the folks at Par Metals too. They're a favorite of the DIY-pro audio commmunity. The 20 series looks like it might work for you. www.par-metal.com
The Crystal chip is made by Cirrus Logic now (they bought Crystal a while back) and is much more easily purchased as a TI PGA2311. The PGA2320 might be a better option, given it's better matching, lower THD, and higher voltage rails (+/-15VDC, vs the 2311s and CS3310s +/-5V)
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
Post Edited (Dave Paton) : 9/25/2005 3:19:49 AM GMT
·· Thanks for the information.· I think I will order the enclosures next.· As for the replacement chip, I noticed right away that it's SOIC.· It will be interesting to see me get into surface mount components finally.· I think to start I will try the CS3310, but I am ordering the other one for later use.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
My $0.02
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· Your advice is taken...Tonight when I get home I will sample that part and start on a design for the PCB.· I may want to find at least one of the old part in DIP format to experiment with first though so I can more easily tweak the circuit prior to committing to PCB.
·· Hey, speaking of the LM12...Did you ever find out why yours developed the hum?· I wondered after looking at the schematic if it was related to a faulty part on your bridge on the output.· That's the only place I could think of from just looking at the schematic.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
There are some versions of the PGA parts in DIP IIRC. The CS3310 only ever came in a SOIC.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
·· Don't know how I missed that on their site.· I am there now and ordering up some samples.· Bummer on the hum.· =(
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
The Phase Linear 700 II, Crown MT1200, Crown D75, Crown K2, and Lab.Gruppen fp6400 all make up for the loss of the LM12 amp. I may resurrect it, sorta, in a pair of powered speakers I'm designing. Then again, I might just use one of the (many) amps that's kicking around and collecting dust. We'll see.
Yes, my house does have twist-loks for the stereo. I know I'm crazy.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
Sounds good. My inbox is always open
I'm looking forward to seeing what you do with your project. Mine is progressing slowly, but a visit from the friendly Noritake rep left me with some amazing samples of some VFDs and I'm looking forward to working on it more and more.
-dave
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
This is not a sig. This is a duck. Quack.