Tom, can we do this to a dual op-amp that has already been bridged? Now, that would be fascinating.
Okay.
Local snubbing bad, remote bypassing bad.
Remote snubbing good, local bypassing good.
Right?
The power supply seems fine except for both common mode noise propagation and missing pi filter section. These appears to be a case of omitted/missing parts. Help? There's many examples, but which ones are the good examples to look at?
As for the preamp bridging module, it appears to be necessary to either use single rail and/or rail splitter and/or the shunt type regulator that Jan posted in the newsletter. Basically, must Avoid speaker return to preamp ground.
For supplementary global feedback, OP275 or AD8620 need to be inverting and bridged, apparently. Well, I cannot figure that out.
Want to listen to your collection without so much urge to skip tracks frequently?
The main problem illustrated:
Documenting the "loudness wars" Hot Mastering (clipping) problem of today:
Rolling Stone article "Death of High Fidelity" <----link
A ready solution:
You can't remaster MP3 without degrading it, so for on-the-fly de-clip of MP3, here's the materials:
De-Clip Care Pack <----link
A no cost high speed de-clipper, such as that kit, turns compressed clipped sound into soft clipped sound, almost like a tube, but with the huge exception that only the clipped part is modified rather than the entire signal. The care pack is Winamp, ffmpg, DX, VST and Foobar2000 compatible. Even an old computer can handle this job.
Alternatively, there are also better quality re-mastering tools available (elsewhere and at a price) to de-clip CD, WAV and FLAC, which can then be saved, since they are theoretically lossless.
Amazingly, if you turn down the noise of hot mastering loudness wars (popular hard clipped studio compression), then after reducing that noise, lowest pitched bass is more apparent.
You're in for a surprise.
The point is that you probably want to reduce the hot mastering noise problem prior to using any sort of dynamic expansion, no matter if that's software or done in live time with a trick amplifier. In either case, you sure don't want the quiet parts quieter and the loud parts louder until after the loud parts are useful quality signal.
The main problem illustrated:
Documenting the "loudness wars" Hot Mastering (clipping) problem of today:
Rolling Stone article "Death of High Fidelity" <----link
A ready solution:
You can't remaster MP3 without degrading it, so for on-the-fly de-clip of MP3, here's the materials:
De-Clip Care Pack <----link
A no cost high speed de-clipper, such as that kit, turns compressed clipped sound into soft clipped sound, almost like a tube, but with the huge exception that only the clipped part is modified rather than the entire signal. The care pack is Winamp, ffmpg, DX, VST and Foobar2000 compatible. Even an old computer can handle this job.
Alternatively, there are also better quality re-mastering tools available (elsewhere and at a price) to de-clip CD, WAV and FLAC, which can then be saved, since they are theoretically lossless.
Amazingly, if you turn down the noise of hot mastering loudness wars (popular hard clipped studio compression), then after reducing that noise, lowest pitched bass is more apparent.
You're in for a surprise.
The point is that you probably want to reduce the hot mastering noise problem prior to using any sort of dynamic expansion, no matter if that's software or done in live time with a trick amplifier. In either case, you sure don't want the quiet parts quieter and the loud parts louder until after the loud parts are useful quality signal.
Last edited:
Step 2 for a useful source. . . a cost effective wonder
There are several options for Audiophile capacity from a computer, and this one is about $16. Skip to the photo (Via Tremor chip) if you're in a hurry.
You may not need to change your computer. Please check the computer with RightMark Audio Analyzer, RMAA, which is free.
The main problem:
Your onboard sound chip works as well as the motherboard power circuit allows and X-Fi fares no better, if when the main problem is noisy motherboard.
Here, when faced with power problems, the Via Envy/Tremor cards take the lead for clean playback, and this one is certainly cheaper than replacing both motherboard and power supply.
Pictured is the cheap version of an expensive success, M-Audio, and it plays almost as well as an M-Audio since Via Envy/Tremor are same design. The upgrade parts pictured are low impedance electrolytic of 0.47uF, put in parallel with the card's output caps to level the high frequency response almost perfectly, so don't even think about using polypro there. In the simple driver software, you can lock it onto 41.1k, which is native for music files and without the destruction of conversions. It can play 16, 24, and 36 bit, so there's no need to hinder MP3's by truncation, uprate, downrate or dithering. It can outperform some specialty DAC products, beating every last one that relies on conversions. Via Tremor/Envy plays at native rates, without added distortion.
After the quick and easy upgrade pictured, the card is successful with RMAA, Classical and cheesecore too. The demands are all different, but they all work clearly. The dynamic impact is prodigious and the chip lives up to its name.
I think that you deserve a well running computer with a spendy audiophile card in it; but, if that's not going to happen any time soon, here's the next best thing. . .
There are several options for Audiophile capacity from a computer, and this one is about $16. Skip to the photo (Via Tremor chip) if you're in a hurry.
You may not need to change your computer. Please check the computer with RightMark Audio Analyzer, RMAA, which is free.
The main problem:
Your onboard sound chip works as well as the motherboard power circuit allows and X-Fi fares no better, if when the main problem is noisy motherboard.
Here, when faced with power problems, the Via Envy/Tremor cards take the lead for clean playback, and this one is certainly cheaper than replacing both motherboard and power supply.
Pictured is the cheap version of an expensive success, M-Audio, and it plays almost as well as an M-Audio since Via Envy/Tremor are same design. The upgrade parts pictured are low impedance electrolytic of 0.47uF, put in parallel with the card's output caps to level the high frequency response almost perfectly, so don't even think about using polypro there. In the simple driver software, you can lock it onto 41.1k, which is native for music files and without the destruction of conversions. It can play 16, 24, and 36 bit, so there's no need to hinder MP3's by truncation, uprate, downrate or dithering. It can outperform some specialty DAC products, beating every last one that relies on conversions. Via Tremor/Envy plays at native rates, without added distortion.
After the quick and easy upgrade pictured, the card is successful with RMAA, Classical and cheesecore too. The demands are all different, but they all work clearly. The dynamic impact is prodigious and the chip lives up to its name.
I think that you deserve a well running computer with a spendy audiophile card in it; but, if that's not going to happen any time soon, here's the next best thing. . .
Attachments
Step 3, a fore-aft balance control regardless of the number of speakers
For right-left, go see Rod Elliot's better balance control. For fore-aft, even if just one speaker (or more), try my "Impedance Thing" circuit in this post.
Pictured here is a way to alter the audible effects of impedance between source and amplifier.
This isn't active powered, doesn't block DC, doesn't shunt, isn't a volume control, doesn't cost very much and doesn't do very much. . . other than exactly what its supposed to do. Basically, it will fake the sound of an input transformer that you didn't have to buy, but still sound like you did.
Other members of "passive buffer" family include Lightspeed Attenuator, input transformer, and some specialty circuits. The audio effect is the same and adjusts the sound field aft, avoiding the "forwards" effect of solid state. Rather than a fixed effect, this one has a dial so that the choice is yours, as for how much. One position of the dial is conveniently Zero Effect (off). All effects should have an off/cancel position.
I'm sure that you will use it at a position close to zero (off), but not all the way to zero. How forwards do you like? That is certainly your choice, just dial it up (fore position is closest to zero-off). Got listening fatigue? No problem! Just dial it down (relaxed position is dialed slightly away from zero-off).
Operation:
You turn the dial until the soundfield recedes aft, and then you turn the dial back to reverse that until zero audible distortion. After several reiterations of this, you have successfully balanced the soundfield fore-aft. Try NOT to adjust it for easily audible effect. Barely audible is okay, but missing the point. The point is perfect fore-aft balance without audible distortion.
Since 3/4ths of the dial travel is useless, you'll want a large knob so that you can be selective for the 1/4 travel that does the intended job.
Parts:
An "A20k" dual gang potentiometer.
A large "Low Impedance Capacitor" in the range of 47uF~3300uF
A small "Low Loss Capacitor" in the range of 100nF~10uF
Other way to say it is "Low ESR"
Possible: 47uF//100nF
Possible: 3300uF//10uF
Anything that's big enough is fine as long as its clear--clarity is the main point, so interview a few likely candidates anywhere in that range. A small amount of experimentation is encouraged for this device. The ground cable follows closely but makes no electrical connection to the potentiometer. This isn't your volume control.
Hookup is in series between source and amp.
See construction examples pictured:
For right-left, go see Rod Elliot's better balance control. For fore-aft, even if just one speaker (or more), try my "Impedance Thing" circuit in this post.
Pictured here is a way to alter the audible effects of impedance between source and amplifier.
This isn't active powered, doesn't block DC, doesn't shunt, isn't a volume control, doesn't cost very much and doesn't do very much. . . other than exactly what its supposed to do. Basically, it will fake the sound of an input transformer that you didn't have to buy, but still sound like you did.
Other members of "passive buffer" family include Lightspeed Attenuator, input transformer, and some specialty circuits. The audio effect is the same and adjusts the sound field aft, avoiding the "forwards" effect of solid state. Rather than a fixed effect, this one has a dial so that the choice is yours, as for how much. One position of the dial is conveniently Zero Effect (off). All effects should have an off/cancel position.
I'm sure that you will use it at a position close to zero (off), but not all the way to zero. How forwards do you like? That is certainly your choice, just dial it up (fore position is closest to zero-off). Got listening fatigue? No problem! Just dial it down (relaxed position is dialed slightly away from zero-off).
Operation:
You turn the dial until the soundfield recedes aft, and then you turn the dial back to reverse that until zero audible distortion. After several reiterations of this, you have successfully balanced the soundfield fore-aft. Try NOT to adjust it for easily audible effect. Barely audible is okay, but missing the point. The point is perfect fore-aft balance without audible distortion.
Since 3/4ths of the dial travel is useless, you'll want a large knob so that you can be selective for the 1/4 travel that does the intended job.
Parts:
An "A20k" dual gang potentiometer.
A large "Low Impedance Capacitor" in the range of 47uF~3300uF
A small "Low Loss Capacitor" in the range of 100nF~10uF
Other way to say it is "Low ESR"
Possible: 47uF//100nF
Possible: 3300uF//10uF
Anything that's big enough is fine as long as its clear--clarity is the main point, so interview a few likely candidates anywhere in that range. A small amount of experimentation is encouraged for this device. The ground cable follows closely but makes no electrical connection to the potentiometer. This isn't your volume control.
Hookup is in series between source and amp.
See construction examples pictured:
Attachments
Well, the accessories posted previously can work with most audio systems.
I work week on week off at two different locations, so that slows me down, since all of personal life and all of my projects are squeezed into every other week. Your question is big size, so in order to avoid too many chapters, I think that you want something doable today. That is workable if you have 8 ohm speakers.
Power:
The revised power supply works as intended; however, a simplistic fusion "best of both" of the two power supply designs would provide better performance. A dual secondaries power supply as in the first example, but with the LED as in the second example, has all of the desired features in one unit. This time, we'd use 4.7k for the led's to facilitate 1/4w resistors and ordinary LED's (preferably Amber or any 2v). AND, we'd also add 6.8k supplementary drainer resistors (totally separately from the LED's although a nearby location is fine). Combined drainer is 2.7k. After unplug, it is enough to eventually pull down that super powerful dual supply.
Amplifier:
If you have 8 ohm speakers a 2 chip parallel build is far easier to construct than a triple and you can still bridge it. This parallel amp takes only minutes. Okay, go look at the amplifier photo posted earlier, and then we downscale to 2 chips, very much easier. You'd have a 470uF cap on the left at v- and a 470uF cap on the right at v+ and you'd have a 100nF cap on the right at v- and a 100nF cap on the left at v+, thereby forming a rigid assembly with the rails held firmly across. This rigid assembly resists spinning off from insulator pads.
It looks like:
470uF------100nF
--------0v-------
100nF------470uF
The compact layout compact can minimize the differences. That is a minimized parts count example and of course you can install more caps in addition to that if you wish. Low impedance, low esr, caps are recommendable for any in direct contact with the amplifier.
The easy way is plan to locate the chips side by side and then start by bending the v-, center pin almost to the face of the chip and then bend the v+ right pin half that far . . . with the Wire Bender (bead work) pliers from arts and crafts, a rounded needle nose pliers (at the chip, hold pin with rounded pliers so it doesn't break off, push pin with finger, avoids stress riser marks on pins). Some 35v caps would provide an easier fit than my 50v pictured.
A pack of 1/8w carbon film resistors would be extremely handy for this project, but especially for the feedback resistor, since you can tie it (wrap once around) to the pins prior to soldering in order to prevent infirm solder connections. There is such low current, any resistor can work at small signal, but do use your digital multimeter to get matched parts.
The triple version is a proof of concept amp. The 2-chip parallel version is ready to go with LM1875, LM675, TDA2030, FAKE LM1875 (at lower voltage!), TDA2040, TDA2050, TDA7293, TDA7293V, TDA7294, TDA7294H. For brighter clearer tone, those not made by National Semiconductor might do better with 220uF or 330uF caps to replace the 470uF caps shown above. Some of the lesser performers may prefer lower gain with lower resistor values, such as 68k/2.7k or 47k/2.2k and these values do make some differences that you might like.
Bridging:
With a 2 chip parallel, you can do an X-bridge (whetstone bridge?). I haven't tried it yet, but it is in the textbooks. Apparently, you take 4 resistors and then form an X with them. Next, the chips are set with the pins of one pair facing the pins of the other pair. For one pair the function is similar to output resistors and for the other pair the function is a resistor mixer feeding the inverting inputs. I have not yet figured out the details on that one. Like this, except for the simple part Simplest Ever Bridging Adapter for Amplifiers <link
Here is a low noise option that is ready to go!:
http://www.diyaudio.com/forums/solid-state/200325-fet-based-phase-splitter.html#post2779768 <link The OP275 jfet or bifet example is similar to Rod Elliot's example and both beat the DRV134 on practical audio performance with random power amplifiers. Bridging Adapter For Power Amps <link These are readily available and produce real results, especially on regulated power. There are TL082 (fits Rod Elliot's design) and regulators at the radio shack. So if you're in a hurry, you really could pick up a good quality phase inverter at the shopping mall, right next to Sears.
I work week on week off at two different locations, so that slows me down, since all of personal life and all of my projects are squeezed into every other week. Your question is big size, so in order to avoid too many chapters, I think that you want something doable today. That is workable if you have 8 ohm speakers.
Power:
The revised power supply works as intended; however, a simplistic fusion "best of both" of the two power supply designs would provide better performance. A dual secondaries power supply as in the first example, but with the LED as in the second example, has all of the desired features in one unit. This time, we'd use 4.7k for the led's to facilitate 1/4w resistors and ordinary LED's (preferably Amber or any 2v). AND, we'd also add 6.8k supplementary drainer resistors (totally separately from the LED's although a nearby location is fine). Combined drainer is 2.7k. After unplug, it is enough to eventually pull down that super powerful dual supply.
Amplifier:
If you have 8 ohm speakers a 2 chip parallel build is far easier to construct than a triple and you can still bridge it. This parallel amp takes only minutes. Okay, go look at the amplifier photo posted earlier, and then we downscale to 2 chips, very much easier. You'd have a 470uF cap on the left at v- and a 470uF cap on the right at v+ and you'd have a 100nF cap on the right at v- and a 100nF cap on the left at v+, thereby forming a rigid assembly with the rails held firmly across. This rigid assembly resists spinning off from insulator pads.
It looks like:
470uF------100nF
--------0v-------
100nF------470uF
The compact layout compact can minimize the differences. That is a minimized parts count example and of course you can install more caps in addition to that if you wish. Low impedance, low esr, caps are recommendable for any in direct contact with the amplifier.
The easy way is plan to locate the chips side by side and then start by bending the v-, center pin almost to the face of the chip and then bend the v+ right pin half that far . . . with the Wire Bender (bead work) pliers from arts and crafts, a rounded needle nose pliers (at the chip, hold pin with rounded pliers so it doesn't break off, push pin with finger, avoids stress riser marks on pins). Some 35v caps would provide an easier fit than my 50v pictured.
A pack of 1/8w carbon film resistors would be extremely handy for this project, but especially for the feedback resistor, since you can tie it (wrap once around) to the pins prior to soldering in order to prevent infirm solder connections. There is such low current, any resistor can work at small signal, but do use your digital multimeter to get matched parts.
The triple version is a proof of concept amp. The 2-chip parallel version is ready to go with LM1875, LM675, TDA2030, FAKE LM1875 (at lower voltage!), TDA2040, TDA2050, TDA7293, TDA7293V, TDA7294, TDA7294H. For brighter clearer tone, those not made by National Semiconductor might do better with 220uF or 330uF caps to replace the 470uF caps shown above. Some of the lesser performers may prefer lower gain with lower resistor values, such as 68k/2.7k or 47k/2.2k and these values do make some differences that you might like.
Bridging:
With a 2 chip parallel, you can do an X-bridge (whetstone bridge?). I haven't tried it yet, but it is in the textbooks. Apparently, you take 4 resistors and then form an X with them. Next, the chips are set with the pins of one pair facing the pins of the other pair. For one pair the function is similar to output resistors and for the other pair the function is a resistor mixer feeding the inverting inputs. I have not yet figured out the details on that one. Like this, except for the simple part Simplest Ever Bridging Adapter for Amplifiers <link
Here is a low noise option that is ready to go!:
http://www.diyaudio.com/forums/solid-state/200325-fet-based-phase-splitter.html#post2779768 <link The OP275 jfet or bifet example is similar to Rod Elliot's example and both beat the DRV134 on practical audio performance with random power amplifiers. Bridging Adapter For Power Amps <link These are readily available and produce real results, especially on regulated power. There are TL082 (fits Rod Elliot's design) and regulators at the radio shack. So if you're in a hurry, you really could pick up a good quality phase inverter at the shopping mall, right next to Sears.
thanks for the response there mate - much appreciated
i've actually PCB'd a triple in Eagle - i'll post a pic up here for your reference and would appreciate any feedback (no pun intended)
re: bridging, I have an LM133 based bridge (based on a Silicon Chip (Australian Magazine) to do the balancing piece. yes i am considering it but want to ensure that the triple setup is fit for purpose, and, can be heatsinked accordingly. and yes the parts for the LM133 bridge are available locally for me (at Jaycar - I am in AU)
i will have a look at the links you've sent though - very much appreciated
i truly appreciate the sound of the LM1875 over the LM3886, TDA2050 and LM3876 I've listened to (as well as other FET based consumer systems) as it has very good 'stage' (nothing like my recently built 12AT7/6L6GC monoblocks i've just pieced together)
i've also started a discussion here where I normally 'lurk' : DIY Audio Projects Forum • Triple Parallel LM1875 amplifier
many thanks for your valuable input - much appreciated
i've actually PCB'd a triple in Eagle - i'll post a pic up here for your reference and would appreciate any feedback (no pun intended)
re: bridging, I have an LM133 based bridge (based on a Silicon Chip (Australian Magazine) to do the balancing piece. yes i am considering it but want to ensure that the triple setup is fit for purpose, and, can be heatsinked accordingly. and yes the parts for the LM133 bridge are available locally for me (at Jaycar - I am in AU)
i will have a look at the links you've sent though - very much appreciated
i truly appreciate the sound of the LM1875 over the LM3886, TDA2050 and LM3876 I've listened to (as well as other FET based consumer systems) as it has very good 'stage' (nothing like my recently built 12AT7/6L6GC monoblocks i've just pieced together)
i've also started a discussion here where I normally 'lurk' : DIY Audio Projects Forum • Triple Parallel LM1875 amplifier
many thanks for your valuable input - much appreciated
I am currently researching the methods for bridging, and welcome input on that topic, especially well working examples that you've used before.
The LM1875 amplifier has a great depth of sound stage because of simple clarity, but it does not have a luxurious width. It is possible that some methods of bridging can widen the sound stage (of each individual speaker even if there is only one speaker total), but it looks like process of elimination to find it.
The way such things go is after the manner of a proper hybrid amp with a triode at input, pushing a solid state output, and whatever is at the input gets amplified. Although I am using all solid state parts, we will have to select a bridging adapter that has its own wide sound stage. . . and then LM1875 will faithfully amplify.
The LM1875 amplifier has a great depth of sound stage because of simple clarity, but it does not have a luxurious width. It is possible that some methods of bridging can widen the sound stage (of each individual speaker even if there is only one speaker total), but it looks like process of elimination to find it.
The way such things go is after the manner of a proper hybrid amp with a triode at input, pushing a solid state output, and whatever is at the input gets amplified. Although I am using all solid state parts, we will have to select a bridging adapter that has its own wide sound stage. . . and then LM1875 will faithfully amplify.
Last edited:
Heat Spreader needed if 4 ohm speakers and bridging.
Bridged Triple
To heat sink triple parallel LM1875 (or 2-chip parallel TDA7294H) to run bridged to a 4 ohm load, use heat spreader method just like an STK chip has built in. You need some thick aluminum (or copper) bar stock from the hardware store (to reach all the way across the heatsink) and some little screws suited to countersink for attaching the chips to the bar. The chips make direct contact to the aluminum bar (except for a spread thin Arctic Ceramique). The kapton/silicon/mica and the shoulder washers are used to electrically isolate the aluminum bar away from the heatsink. This use is at capacity, so its good to consider means of current limiting (like PI filter overdo, fuses, cable trick, etc) for adding durability. This also needs a well ventilated enclosure with air intake at bottom and air output at or near top. Generic chips cannot withstand this load, and even authentic chips should have current limiting added.
Bridged Triple
For driving 6 ohm speakers, you can use a simple mica or silicon sheet.
Bridged double
For driving 8 ohm speakers, you can use 2-chip parallel and a simple mica or silicon sheet.
Non-bridged and easy
The 2-chip parallel amp will also run 4 ohm speakers to 50 watts if you don't bridge, and this also uses a simple mica or silicon sheet.
Bridged Triple
To heat sink triple parallel LM1875 (or 2-chip parallel TDA7294H) to run bridged to a 4 ohm load, use heat spreader method just like an STK chip has built in. You need some thick aluminum (or copper) bar stock from the hardware store (to reach all the way across the heatsink) and some little screws suited to countersink for attaching the chips to the bar. The chips make direct contact to the aluminum bar (except for a spread thin Arctic Ceramique). The kapton/silicon/mica and the shoulder washers are used to electrically isolate the aluminum bar away from the heatsink. This use is at capacity, so its good to consider means of current limiting (like PI filter overdo, fuses, cable trick, etc) for adding durability. This also needs a well ventilated enclosure with air intake at bottom and air output at or near top. Generic chips cannot withstand this load, and even authentic chips should have current limiting added.
Bridged Triple
For driving 6 ohm speakers, you can use a simple mica or silicon sheet.
Bridged double
For driving 8 ohm speakers, you can use 2-chip parallel and a simple mica or silicon sheet.
Non-bridged and easy
The 2-chip parallel amp will also run 4 ohm speakers to 50 watts if you don't bridge, and this also uses a simple mica or silicon sheet.
Last edited:
Love your work.
OK bridging - you've obviously looked at the DRV134 - a very simple setup for bridging.
The LM133 setup can be found here on SiliconChip - Silicon Chip Online - Balanced/unbalanced converter for audio work if you click on the picture you can look at the other pictures in the article also.
there is also a rudimentary setup with NE5532 (i have a schematic somewhere) - but boards are available for this on fleabay. i believe the NE5532 chip can be 'rolled' of sorts but have not gone anywhere with this (yet) http://www.diyaudio.com/forums/digital-line-level/187147-drop-replacement-ne5532.html
hope this helps you in some way mate.
OK bridging - you've obviously looked at the DRV134 - a very simple setup for bridging.
The LM133 setup can be found here on SiliconChip - Silicon Chip Online - Balanced/unbalanced converter for audio work if you click on the picture you can look at the other pictures in the article also.
there is also a rudimentary setup with NE5532 (i have a schematic somewhere) - but boards are available for this on fleabay. i believe the NE5532 chip can be 'rolled' of sorts but have not gone anywhere with this (yet) http://www.diyaudio.com/forums/digital-line-level/187147-drop-replacement-ne5532.html
hope this helps you in some way mate.
Cool - I've seen the copper bar method employed on some high end LM3886 amplifiers. My heatsink is as follows, each LM will be mounted directly to the heatsink insulated by silicon sheet and the screw also insulated to ensure no connectivity. The key I have found thusfar is when soldering the LM1875 to the PCB, mount the LM's to the Heatsink, then align the pins of the LM to the PCB, and then solder in place. This ensures maximum contact to the heatsink.
Here is my heatsink : http://jaycar.com.au/ShowLargephoto.asp?id=703&PRODNAME=Diecast Heatsink - 150mm Long&IMAGE=
Its thermal resistance is 0.784°C/W.
Its 150mm long and 75mm high (6 inch by 3 inch) and my triple PCB is 100 x 100
Here is my heatsink : http://jaycar.com.au/ShowLargephoto.asp?id=703&PRODNAME=Diecast Heatsink - 150mm Long&IMAGE=
Its thermal resistance is 0.784°C/W.
Its 150mm long and 75mm high (6 inch by 3 inch) and my triple PCB is 100 x 100
Thanks man!
I think we need to add a current limiting method to that. There's a lot of ways to current limit a power amplifier. Thank goodness LM1875 doesn't have the screechy spike system; however, we probably do need to add our own low noise provision for current limiting. As a bonus, that could increase the RMS output slightly by softening the clipping performance, aka retard power available for clipping right on time. PI filter overdo (bigger R value in the CRC) is the popular method for guitar amplifiers. Ribbon cable to the transformer is the popular method in retail amplifiers. Surely we can think up something and then add fuses just in case. I like mains fuses and speaker jack fuses. Neither one of us are shy of non-zero output impedance and so speaker jack fuses should be okay.
I think we need to add a current limiting method to that. There's a lot of ways to current limit a power amplifier. Thank goodness LM1875 doesn't have the screechy spike system; however, we probably do need to add our own low noise provision for current limiting. As a bonus, that could increase the RMS output slightly by softening the clipping performance, aka retard power available for clipping right on time. PI filter overdo (bigger R value in the CRC) is the popular method for guitar amplifiers. Ribbon cable to the transformer is the popular method in retail amplifiers. Surely we can think up something and then add fuses just in case.
I like mains fuses and speaker jack fuses. Neither one of us are shy of non-zero output impedance and so speaker jack fuses should be okay.
Last edited:
The circuit board can also be equipped with mounting holes whereby the circuit board can be attached for the purpose of further pressing the LM1875 against the heatsink a bit.
By all means feel free to post your triple LM1875 project photos here if you like. Chip amp design mainly varies by layout and a few resistors. It would be great to work together on this project. I'm sure that many people would rather use a circuit board for the triple parallel amplifier.
pleasure.
from a simple perspective, you need to determine the behaviour when clipping occurs, then, then shunt current to a null device.
theres quite a good article here : Power Amplifier Clipping
from a simple perspective, you need to determine the behaviour when clipping occurs, then, then shunt current to a null device.
theres quite a good article here : Power Amplifier Clipping
First, let's try using a 2.2 ohm resistor to create a separate signal star ground, where the NFB cap(s) the input load and the input ground all find their ground reference. Whether we actually use the "ground lift resistor" or not, it will still alter the layout to keep power off of small signal.
Second, 100nF caps are usually quite a bit smaller, and so there needs to be an additional pad to support the normal size.
Third, the spacing between the LM1875 and the edge of the board is a bit far, sort of hard to flex the chip that much, so perhaps move closer to the edge of the board.
Fourth, apparently the output resistors, large 0.2uF, have vanished and been replaced by a dramatic feed forward system? It appears to be unique. I'm just not familiar with that system, so, please review the red speaker output cable/trace and see what is going on there.
Fifth, The 2uF input cap needs an additional pad to support Nichicon ES 1uF which is a rather small electrolytic of the finest quality, for only a few cents,and I think that we should at least be able to use it or an Elna New Cerafine//tiny polyester for a reference point. Both good and bad caps are available in all sizes.
P.S.
I think that you could look at some of the BPA300 and PA150 boards, because the pinouts and layout for LM1875 is not a great deal different from LM3886.
Second, 100nF caps are usually quite a bit smaller, and so there needs to be an additional pad to support the normal size.
Third, the spacing between the LM1875 and the edge of the board is a bit far, sort of hard to flex the chip that much, so perhaps move closer to the edge of the board.
Fourth, apparently the output resistors, large 0.2uF, have vanished and been replaced by a dramatic feed forward system? It appears to be unique. I'm just not familiar with that system, so, please review the red speaker output cable/trace and see what is going on there.
Fifth, The 2uF input cap needs an additional pad to support Nichicon ES 1uF which is a rather small electrolytic of the finest quality, for only a few cents,and I think that we should at least be able to use it or an Elna New Cerafine//tiny polyester for a reference point. Both good and bad caps are available in all sizes.
P.S.
I think that you could look at some of the BPA300 and PA150 boards, because the pinouts and layout for LM1875 is not a great deal different from LM3886.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.